New Perspectives on Current Development Policy: Covid-19, the Digital Divide, and State Internet Regulation (SpringerBriefs in Economics) 3030884961, 9783030884963

Citation preview

SPRINGER BRIEFS IN ECONOMICS

Jeffrey James

New Perspectives on Current Development Policy Covid-19, the Digital Divide, and State Internet Regulation 123

SpringerBriefs in Economics

SpringerBriefs present concise summaries of cutting-edge research and practical applications across a wide spectrum of fields. Featuring compact volumes of 50 to 125 pages, the series covers a range of content from professional to academic. Typical topics might include: • A timely report of state-of-the art analytical techniques • A bridge between new research results, as published in journal articles, and a contextual literature review • A snapshot of a hot or emerging topic • An in-depth case study or clinical example • A presentation of core concepts that students must understand in order to make independent contributions SpringerBriefs in Economics showcase emerging theory, empirical research, and practical application in microeconomics, macroeconomics, economic policy, public finance, econometrics, regional science, and related fields, from a global author community. Briefs are characterized by fast, global electronic dissemination, standard publishing contracts, standardized manuscript preparation and formatting guidelines, and expedited production schedules.

More information about this series at http://www.springer.com/series/8876

Jeffrey James

New Perspectives on Current Development Policy Covid-19, the Digital Divide, and State Internet Regulation

Jeffrey James Tilburg University Tilburg, The Netherlands

ISSN 2191-5504 ISSN 2191-5512 (electronic) SpringerBriefs in Economics ISBN 978-3-030-88496-3 ISBN 978-3-030-88497-0 (eBook) https://doi.org/10.1007/978-3-030-88497-0 © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Acknowledgements

As with my previous brief for Springer, my primary debt is to Josette Janssen and Niko Chtouris. Josette again played the multiple roles of typist, editor, and organizer in her unique way. I benefitted greatly from her perceptive observations and her ability to somehow keep track of all the different versions of the chapters, and I sincerely do not know what I would have done without her. A sizeable debt is owed again to Niko Chtouris of Springer, who is one of the very best editors I have worked with over the years. Not only is he quick to see creative possibilities, but he is always encouraging and amazingly quick to respond to queries and problems. It is comforting to feel that an editor is a straight shooter, rather than someone who will spring surprises very late in the process, which has happened to me on more than one occasion. Before turning to specific copyright permissions, I need to first acknowledge my debt to the ideas promulgated in the final chapter of my last book, entitled ‘The Impact of Smart Feature Phones on Development: Internet, Literacy and Digital Skills’, 2020, and to parts of the paper written by myself and Padraig Carmody, entitled ‘The Global Digital Divide-Reversals, Leaps and Paradoxes: Interrogating the Potential of Digital Technology in the Global South’ (under review, 2021). In no particular order, I am grateful to those who have granted me permission to use their copyright material. The World Bank for permission to use Tables 3.1, 3.3, 3.4, and 3.5 and 5.3, 5.8, and 5.14 under license CC-BY-3.0-IGO or CC-BY-4.0; Statista for permission to use Tables 3.6, 5.2, and 5.10; the World Economic Forum for permission to use Tables 3.5, 3.8, 4.1, 5.3, 5.5, 5.13, 6.1, and 6.2, under the Creative Commons Attribution-Non-Commercial-No Derivatives 4.0 International Public License; YouGov/Imperial College of London surveys from 22 to 25 June 2020 for permission to use Tables 2.1 and 2.2; GSMA Intelligence for permission to use Tables 5.1, 5.4, 5.6, 5.7, 5.8, 5.9, and 5.12; ITU for permission to use Table 3.2; CIA World Factbook 2018 for permission to use Table 3.3, which is in the public

v

vi

Acknowledgements

domain; and Alliance for Affordable Internet (2020) for permission to use Table 5.11 from Luxury to Lifeline: Reducing the Cost of Mobile Devices to Reach Universal Internet Access. Web Foundation.

Contents

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Main Conclusions by Chapter . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . .

. . . .

1 1 6 7

2

Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in Developing Countries . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Resolution of Cognitive Dissonance . . . . . . . . . . . . . . . . . . . . . . . 2.3 Dissonance Theory and the Pandemic . . . . . . . . . . . . . . . . . . . . . 2.4 Policy Derived from Dissonance Theory . . . . . . . . . . . . . . . . . . . 2.4.1 Forced Compliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Non-Coercive Methods of Compliance . . . . . . . . . . . . . . . . . . . . . 2.5.1 Group Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.2 The Foot-in-the-Door Technique . . . . . . . . . . . . . . . . . . . . 2.5.3 Induced Hypocrisy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.4 The Dynamics of Dissonance Policy . . . . . . . . . . . . . . . . . 2.5.5 Nudging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

9 9 10 10 12 12 13 13 15 16 18 19 19 21

Digital Divide Reversal: Evidence, Explanations, and Implications . . 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Digital Divide Reversal: The Data . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Explanations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Demographics and the Popularity of Gaming . . . . . . . . . . . 3.3.2 The Differential Value of Time . . . . . . . . . . . . . . . . . . . . . 3.3.3 Communications and Infrastructural Differences . . . . . . . . 3.3.4 Costs of Internet Data and Affordability . . . . . . . . . . . . . . 3.4 Conclusions and Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

23 23 24 27 27 28 29 31 32 34

3

vii

viii

4

5

6

Contents

Why Is India So Dominant in the Demand for New Smart Feature Phones That Are Internet Connected? . . . . . . . . . . . . . . . . . . . . . . . 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Explaining the Indian Experience with the JioPhone in Comparative Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1 Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Pre-Adoption (Binary Decisions) . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Adoption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6 Use of Adopted Smart Feature Phones . . . . . . . . . . . . . . . . . . . . . 4.7 Welfare Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interregional and Intercountry Analysis of Mobile Internet Connectivity in Sub-Saharan Africa . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Interregional Analysis of Determinants of Mobile Internet Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Affordability: Handsets, Data, and Incomes . . . . . . . . . . . . . . . . . 5.4 Digital Skills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 Rural–Urban Gaps by Region . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 Intercountry Variations in Internet Connectivity and its Determinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.1 Affordability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.2 Network Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.3 Digital Skills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mobile Use of the Internet Among the Poor in the Global South: Preferences, Theories, and Policies . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 The Argument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 The Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Alternative Explanations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Digital Skills, Information, and the Choice of Leisure . . . . . . . . . . 6.6 Welfare Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7 Policy Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. 37 . 37 . 38 . . . . . . . .

39 39 41 41 44 45 45 47

. 49 . 49 . . . .

50 51 53 54

. . . . . .

55 56 57 58 60 61

. . . . . . . . . .

63 63 64 66 69 70 71 72 74 75

Chapter 1

Introduction

1.1

Introduction

All the chapters below are concerned with improving policy-making in selected major contemporary issues in developing countries. These include the COVID-19 pandemic; Internet connectivity in sub-Saharan Africa, which despite recent improvements still lags behind nearly all other global regions; why some low-income countries in Africa nevertheless exhibit exceptional performance and what lessons for policy can be drawn therefrom; why the market for the new smart feature phones is so dominated by India; and what can be learnt from that and the fact that poor countries on average spend more time on the Internet than the rich and the implications that are associated with that. These issues were selected for several reasons. One of the most important was that they have mostly not yet received much attention in the literature, for example, the lack of application of insights from social psychology to the COVID-19 pandemic in developing countries. A second reason is that the prevailing policy response to these issues has, in general, been inadequate to the task, and the final one is that the issues themselves are of major importance, most obviously the COVID-19 pandemic, but also the use of the Internet for entertainment rather than development purposes. This, I believe, is partly due to an acute shortage of digital skills in developing countries, which causes a severe waste of economic opportunities in the domestic economy and abroad (as in a thin integration into the world economy). Yet, the attention paid to such skills in much of the developing world can only be described as uneven and inadequate, especially in the isolated rural areas in this part of the world. In terms of method, the chapters that follow differ quite sharply from the typical academic papers based on complex models and multivariate methods. They can be better described as ‘perspective’ pieces, whose primary function is rather to bring new and neglected issues to a wider audience. In fact, because of the novelty of the issues that they discuss, the chapters rely, to a greater extent than is usual, on sources such as industry reports, data from international organizations, and unpublished © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. James, New Perspectives on Current Development Policy, SpringerBriefs in Economics, https://doi.org/10.1007/978-3-030-88497-0_1

1

2

1 Introduction

material. According to one source ‘They [perspective pieces] are different from other types of articles because they present a different take on an existing issue, tackle new and trending issues, or emphasize topics that are important, but have been neglected, in the scholarly literature’ (Enago Academy 2019, emphasis added). In the case of the COVID-19 pandemic discussed in Chap. 2, for example, policymakers have often been loathe to impose coercive measures, even though the prevailing conditions in most developing countries are akin to wartime conditions, when, in say the Second World War, harsh restrictions were imposed on individuals in order to bolster the war effort. Currently, by contrast, governments have too often allowed certain institutions to remain open even in the face of many lost lives.1 In Asia, on the other hand, there has been much greater societal support for unpopular measures. Many of the countries in the region, moreover, were relatively quick to respond to the early onset of what was to become a full-blown pandemic. There is also a policy need to respond to situations where governments seek to alter behaviour in a non-coercive manner (such as with respect to wearing masks, social distancing, and so on). For this category, I apply insights from a well-known theory from social psychology, known as cognitive dissonance. In particular, I am referring to group pressure, the ‘foot-in-the-door’ technique and induced hypocrisy. The idea is to make people do what they do not want to do without forcing them. Chapter 3 is designed to respond to what I refer to as a digital divide ‘reversal’. That is, that despite the typical pattern of the divide, in which the gains from new technology accrue to rich rather than poor countries, an exception seems to occur when the latter appear to gain more because they spend more time on the Internet than the former and the time they do spend is devoted mainly to entertainment rather than development issues (such as searching for a job). Appropriate policy, in this case, however, needs to consider the welfare implications of these findings, rather than simply adopting a laissez-faire attitude to them. More is not necessarily better, if the extra time supplants other more valuable activities such as work on the family farm or (in the case of schoolchildren) doing homework. All too often, in my view, policy is made on the assumption that individuals know what is best for them, an assumption that has been widely criticized in the psychology literature.2 Moreover, there are many cases in developed countries where the state overrides individual choices, as when, for example, seatbelts are made compulsory, as sometimes is education and certain forms of insurance. And the avid use of gaming, especially in Asia, is known to cause addiction, which is now recognized in the USA as a serious medical condition. In

They have also tended to impose minimal fines on those who transgress mandated behaviour. As I note in Chap. 2, however, people are generally very reluctant to give up their long-standing views and behaviour. 2 By Thaler (1992) among others. 1

1.1 Introduction

3

many if not most developing countries, on the other hand, addiction and mental health more generally do not receive the attention they deserve.3 Chapter 4 is concerned with understanding why India is so dominant in the market for smart feature phones and on this basis to elicit policies that may be suitable for other developing countries as well. As in some other chapters, the discussion is concentrated on the main determinants of Internet connectivity, such as broadband coverage, price of handsets and data, literacy, digital skills, and relevance. As already noted, it is policy in each of these areas that determines the overall performance of a region or country. In India, though, it was not so much state policy towards the JioPhone that seems to have been crucial. Rather, it was the heavy subsidization of this product by the country’s largest conglomerate, Reliance Industries (whose behaviour in part was motivated by the goal of eliminating competitors in a price war).4 In fact, in the initial years, the subsidies enabled handsets to be sold for almost nothing. The price war thus initiated by the entry of the JioPhone also contributed to a decline in data costs, which, at one point, were the lowest in the world. And in relation to broadband coverage, Reliance Jio was again involved. Thus, ‘The most significant expansion in mobile broadband coverage in 2019 occurred in India, as Reliance Jio covered almost 99% of the population with its 4G network . . . surpassing both 2G and 3G coverage’ (GSMA 2020, p. 40). These Indian examples point up the potential of the private sector in helping to increase the country’s connectivity and the benefits available from the Internet. In Africa, as well, KaiOS Technologies has partnered with two of the region’s largest telecom operators to bring the smart feature phones to a region that most severely lacks the benefits that Internet can bring, especially to isolated rural localities. In short, though the state needs to play a much more active role in alleviating the constraints to further Internet use (perhaps most notably in relation to digital skills), there is also much that can be done by invoking the help of the private sector. The data cited so far, however, have been pitched at the country level which ignores the (often marked) division into rural and urban areas. This omission is regrettable. In low- and middle-income countries, for example, the rural population is 37% less likely than the urban region to use mobile Internet (GSMA 2020, p. 17). In India, as elsewhere, this is due partly to the rural–urban income gap, which makes mobile Internet access much less affordable to the low-income inhabitants of rural areas. According to scholars such as Lipton (1977), the gap itself is due to an urban bias, which, for mainly political reasons, favours urban areas in the allocation of resources and other policy issues related to education, health, and so on. It follows from this

3

It is notable here that in recent years, gaming disorder has been included by the World Health Organization (WHO) in the International Classification of Diseases (the ICD). This classification is used by medical personnel around the world to diagnose conditions and by those who conduct research to categorize these (WHO 2018). 4 For details on the Jio case, see James (2020).

4

1 Introduction

line of argument that countervailing measures need to be taken in order to reduce or close the digital gap and to bring about a more egalitarian pattern of development, which, in turn, will help to close the divide in Internet connectivity. And in spite of the manifold political and other difficulties involved, it is not impossible. For example, consider the Prime Minister’s Rural Digital Literacy Campaign in India, begun in 2016, with the intention of bringing digital literacy to 60 million rural households by March 2019. For such a purpose, the scheme received in excess of US $350 million. And recall from above that Reliance Jio has reached almost complete 4G coverage of the population in India.5 Digital help is also afforded to rural inhabitants in India and elsewhere, who do manage to gain access to the JioPhone, in the form of an application called ‘Life’, which among many other things provides help with digital skills to first-time users of the Internet.6 The application also provides free health information and knowledge about gender issues. What is perhaps especially important in the rural context is the assistance given to small farmers. In particular, through a combination of localized farming information from ‘iCow, Wikipedia and local sources as well as farmer-tofarmer conversations, the Agriculture section of the Life app helps small farmers improve their businesses’ (Meta 2020). In Africa, a pilot project between ‘Justdiggit’ and KaiOS has recently addressed itself to farmers in Tanzania who suffer degradation of vast swathes of fertile farmlands (due to climate change and misuse of land), which leads to ‘failed harvests, poverty, and climate refugees. In Africa, around 350 million smallholder farmers deal with these issues already or will have to soon’ (Leung 2021, n.p.). To combat land degradation, a Regreen App has been developed for KaiOS Technologies, based on Justdiggit’s scalable regreening techniques, which have regreened 60,000 Ha of (farm) lands in Africa within three years. ‘This effort includes bringing back five million trees, which save 11 billion litres of water annually and capture more than 1.8 million tons of CO2e’ (Leung 2021, n.p.). Chapter 5 is concerned with the determinants of Internet connectivity in sub-Saharan Africa, at both regional and country levels. These include, as before, literacy and digital skills, affordability, and local content/relevance. A comparison of regions along these lines shows that even after rapid recent progress, there is only one case, digital skills, in which the region in question reaches the level of another (region). And even this achievement depends on the particular measure of such skills, which was adopted for the sake of comparison (from the Global Competitiveness Index of the World Economic Forum). What the regional analysis fails to capture, however, is that in sub-Saharan Africa and other developing countries, there will be outlier countries which perform well above average and from which more specific policy conclusions can be drawn. After

5

There are in fact many examples of state intervention in favour of the poor rural areas. Recently in Asia for example, there has been talk of a universal basic income, or a minimum living standard guarantee (SCMP 2017). 6 For a full discussion of this novel app, see Leung (2019).

1.1 Introduction

5

all, policy is mostly made at the country level and that is why the second half of the chapter is devoted to comparisons of countries within sub-Saharan Africa. Overall, I conclude that income plays a major role in determining which countries appear as positive outliers. For example, Mauritius (a high-income country) and South Africa (a high-middle-income nation) occupy the top two places on many determinants of Internet connectivity. This is only to be expected, given the centrality of income in determining coverage, affordability, education, and relevance. Income, however, is not the only factor that explains the pattern of country performance. Kenya and Ghana, for example, are only low-middle-income countries but appear frequently in the tables of outlying countries. The explanation for this seems to be in part that these countries have long maintained a serious policy commitment to technological issues, including those related to policy towards information technology. Both countries, for instance, have a relatively high number of Internet hubs though they do not enjoy relatively high incomes in the region. Rwanda is a country that belongs to the lowest income group, but which has nevertheless made considerable progress in information technology, largely because of a serious and prolonged state commitment to foster this type of technology. Most striking of all, perhaps, is the list of countries with the cheapest data costs. In particular, the two cheapest nations in Table 5.10 are Somalia and Sudan, two of the world’s poorest countries, while the others all belong to the low-middle-income group. What is notable is thus that factors other than income seem to be most important in explaining the outcome. In the case of Somalia, the leading outlier, the reason is said to be the number of telecom operators and the intensification of competition, but in this case, as in many others discussed in the book, far more research needs to be done. A major finding of Chaps. 4 and 5, therefore, is that whereas much can be attributed to income in determining connectivity, there are also areas in which even poor countries can make a notable contribution. These experiences warrant more research than has thus far been devoted to them. This is so because they may offer lessons to other very poor countries, which are usually least able to meet the Sustainable Development Goals. This was the case, for example, with the ‘basic needs’ literature of the 1970s, which, among other things, sought to correlate basic needs (such as education, nutrition, and so on) with satisfaction for countries, against their incomes per capita. While there was a close correlation between the two, there were several important outliers. One was the case of Brazil, with a high income but relatively low basic needs satisfaction, and the other was Sri Lanka, which despite its low income performed better than the correlation predicted. Broadly, the difference lay in contrasting development policies. For, whereas Brazil paid little attention to meeting basic needs, Sri Lanka sought explicitly to meet such needs among a wide section of the population and, indeed, for many years, was regarded as a model for those concerned with promoting egalitarian development among poor countries (Streeten et al. 1981). Chapter 6 is related to Chap. 3 in that both deal with the finding that the poor tend to spend more time on leisure than development Internet uses. The former, however,

6

1 Introduction

deals in particular with the well-known anthropological interpretation of this finding by Arora (2016). I argue that her approach ignores the vast amount of time spent on leisure by the poor, amounting in some cases to three or four hours a day. I argue, furthermore, that she ignores alternative explanations of the data such as that the lack of digital skills effectively precludes the use of more sophisticated developmental activities on the Internet. In my view, a much more international policy is needed than Arora allows.

1.2

Main Conclusions by Chapter

The main conclusions to be drawn, by chapter, are these: Chapter 2: Changing behaviour in a socially desirable direction without coercion can be promoted by means of policies derived from the well-known psychological theory of cognitive dissonance. Moreover, there is a case to be made for more coercive policies than are currently adopted in many developing countries (with the exception of those in Asia). Chapter 3: Considers the response to the finding of a reversal in the digital divide. That is, that contrary to the usual outcome that the benefits accrue to the rich countries, developing countries spend more time on the Internet than the rich and arguably derive greater benefits. On the other hand, the poor spend more time on social media than the rich and much of the entire time spent on the Internet tends to be devoted to entertainment rather than development activities. This leads to a loss of development potential domestically, but also to a ‘thin’ integration into the international economy. One of the most needed policy reforms here is the promotion of digital skills, whose scarcity lies behind the focus on entertainment rather than development. Chapter 4: Is a case study of the JioPhone in India, which has achieved great success there, more indeed than any other country. To explain this, an analysis was conducted of how India performed on the determinants of connectivity such as digital skills, affordability, relevance, and so on. At least in the early years, however, what made India exceptional was the affordability of the handset and data. As regards the former, the JioPhone was designed to meet the needs of the poor rather than the rich, and this lowered the price considerably. But what really mattered was the heavy subsidization of the phone by Reliance Industries, India’s largest conglomerate, which caused the handset to be sold for almost nothing. Moreover, data were available at exceptionally low prices. It remains to be seen though how KaiOS, the software developer, will fare at the higher cost of about US$20 in Africa, where the smart feature phone is being widely introduced (and where most determinants of connectivity assume lower values than in India). Chapter 5: The focus here is on sub-Saharan Africa—how it compares with other regions on the determinants of Internet connectivity and which countries in the region emerge as outliers in comparisons at this level. The main conclusion of the country analysis is that whereas income plays an important explanatory role, there

References

7

are also cases where even very poor countries are seen to render exceptional performance on some determinants (Somalia is one such case). Chapter 6: The last chapter has an affinity with Chap. 3 in that both deal with the finding that the poor in the Global South invest their Internet time on leisure rather than development-oriented activities. The former, however, is specifically devoted to a critique of the anthropological interpretation of the finding by Payal Arora. I argue in favour of a more interventionist approach to the issue than she allows by relying on concepts such as consumer sovereignty.

References Arora PA (2016) The next billion users: digital life beyond the West. Harvard University Press Enago Academy (2019) How to share opinion on research articles. Available https://www.enago. com/academy/perspective-how-to-share-opinion-on-research-articles/ GSMA (2020) The state of mobile internet connectivity. Available https://www.gsma.com/r/wpcontent/uploads/2020/09/GSMA-State-of-Mobile-Internet-Connectivity-Report-2020.pdf James J (2020) The smart feature phone revolution in developing countries: bringing the internet to the bottom of the pyramid. Inf Soc 36(4):226–235 Leung E (2019) Life, a new initiative by KaiOS to help first-time Internet users, KaiOS. Available https://www.kaiostech.com/life-a-new-initiative-by-kaios-to-help-first-time-internet-usersmake-the-most-of-mobile-internet-access/ Leung E (2021) Justdiggit and KaiOS partnership regreens and connects rural Tanzania. Available https://www.kaiostech.com/justdiggit-and-kaios-partnership-regreens-and-connects-ruraltanzania/ Lipton M (1977) Why poor people stay poor: a study of urban bias in world development. Australian National University Press Meta A (2020) How the Life app is changing lives around the world. KaiOS Technologies. Available https://www.kaiostech.com/life-app-changing-lives-around-world/ South China Morning Post (SCMP) (2017) Getting paid to do nothing: why the idea of China’s dibao is catching on. Available https://www.scmp.com/week-asia/article/2087486/getting-paiddo-nothing-why-idea-chinas-dibao-catching Streeten P, Burki SJ, ul Haq M, Hicks N, Stewart F (1981) First things first: Meeting basic human needs in the developing countries. The World Bank-Oxford University Press Thaler R (1992) The winner’s curse: paradox and anomalies in economic life. Free Press World Health Organization (WHO) (2018) WHO releases new international classification of diseases. Available https://www.who.int/news-room/spotlight/international-classification-ofdiseases

Chapter 2

Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in Developing Countries

2.1

Introduction

Probably, the best-known theory in social psychology is cognitive dissonance, and it is concerned with the unpleasant psychic state that is aroused when beliefs and behaviour are out of step with one another. Much attention is paid, for example, to the question of how such inconsistency (dissonance) is dealt with by the individual and under what circumstances. This matters for policy towards the pandemic, as it may or may not work in the desired direction. More generally, this chapter deals with the issue of how dissonance theory can be applied to measures designed to mitigate the rampant and disastrous effects of COVID-19 in all corners of the globe. This is an exercise, which so far as I am aware has not yet been conducted. The policies discussed fall into two broad categories, namely, those that involve widespread coercion of the population and those that induce desired behaviour change without coercion. Practical examples of each type are presented, of success as well as failure. What is especially appealing, from my point of view, is that dissonance theory makes apparently irrational or obdurate behaviour, explicable in psychological terms. And in so doing it offers us the chance to make policy that is consistent with, rather than antagonistic towards, seemingly aberrant human behaviour. In some cases, however, widespread coercion may be needed to save lives and reduce transmission. It is true that while aspects of the relationships between dissonance theory and policy towards the pandemic have been alluded to in the literature, few systematic reviews have yet been undertaken that focus on developing countries but also include some that are developed. According to Smith and Gibson (2020, n.p.), for example, ‘There is very little published behavourial science research on pandemics.’ Moreover, Cooper (2019) has recently argued that ‘many researchers stop short of the goal of turning [their] research into bona fide practices. My suggestion is for dissonance theorists to become more engaged in people’s lives by providing treatments that are available © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. James, New Perspectives on Current Development Policy, SpringerBriefs in Economics, https://doi.org/10.1007/978-3-030-88497-0_2

9

2 Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in. . .

10

for people to use’ (Cooper 2019, n.p.). The purpose of this chapter, accordingly, is to undertake just such an endeavour.1 To begin with, I set forth the basic tenets of the dissonance formulation, as a prelude to a brief discussion of the ways in which the theory explains the recalcitrant response to some current policy efforts to ameliorate the dire health consequences of the pandemic. The main section of this chapter then describes a series of possible policy alternatives to address the main problem thus outlined.

2.2

Resolution of Cognitive Dissonance

It needs to be recognized that there is not just one but two ways in which dissonance can be ameliorated by the conflicted individual. The first is where he or she alter their beliefs so as to make them consistent with their behaviour. The second is where, instead, the latter is made to conform with the former. Dissonance theory, however, is ‘nonspecific with regards to the mode of dissonance reduction used by the individual after dissonance arousal’ (Brehm and Cohen 1962, p. 277). What is said is that in general the mode of dissonance reduction will be chosen according to the one that affords the least resistance. The question is then whether it is easier to alter one’s new behavior or one’s old attitudes. Among the crucial determinants of this choice are, firstly, the degree of irrevocability of the former, e.g., whether or not the individual would incur sizable financial losses from ‘uncommitting’ himself from his new behavioral commitment. The difficulty of changing one’s attitudes—the second major determinant—depends on the one hand, on how entrenched these have become and on the other hand, on how much social support one has in changing them (Brehm and Cohen 1962, p. 290).

2.3

Dissonance Theory and the Pandemic

The clearest description of the role played by dissonance in the pandemic is to be found in a recent article by Aronson and Tavris (2020). In particular, cognitive dissonance is described as ‘the motivational mechanism that underlies the reluctance to admit mistakes or accept scientific findings—even when, those findings can save our lives’ (Aronson and Tavris 2020, n.p.). The authors go on to suggest that ‘This dynamic is playing out during the pandemic among the many people who refuse to wear masks or practice social distancing. Human beings are deeply unwilling to change their minds. And when the facts clash with their preexisting conditions, some people would sooner jeopardize their health and everyone else’s than accept new

1

James and Gutkind (1985) discuss some of the lessons from dissonance theory for policy in developing countries, but well before the coronavirus had emerged. Some of the policies do nevertheless overlap.

2.3 Dissonance Theory and the Pandemic

11

information or admit to being wrong’ (Aronson and Tavris 2020, n.p., emphasis added). From the point of view of policy, moreover, Aronson had earlier brought about an advance in dissonance theory, by ‘demonstrating the powerful, yet nonobvious, role it plays when the concept of self is involved. Dissonance is most painful when evidence strikes at the heart of how we see ourselves—when it threatens our belief that we are kind, ethical, competent, or smart’ (Aronson and Tavris 2020, n.p.). Arguably, it is this intense kind of experience that underlies the behaviour of those who refuse to take protective measures against COVID-19, but who know all too well the life-threatening consequences of such behaviour. Put another way, it is in these cases that dissonance is most acutely felt and especially difficult to resolve. What adds to the difficulty, moreover, is that the theory appears to exhibit a pernicious dynamic, namely, that soon after a decision is made, the individual begins to rationalize it, by, for example, stressing the advantages of his choice and exaggerating the weak points of the alternative, rejected options. What this effect implies for policy is that efforts to effect a change in behaviour towards COVID-19 prevention will become increasingly difficult over time as the errant behaviour becomes more and more entrenched (see more on this below).2 Indeed, as Aronson and Tavris (2020, n.p.) describe the solidification behaviour over time, ‘Before long, any ambivalence we might have felt at the time of the original decision will have morphed into certainty. As people justify each step taken after the original decision, they will find it harder to admit they were wrong at the outset. Especially when the end result proves self-defeating, wrongheaded, or harmful’. Applied to the coronavirus, this strand of dissonance theory leads to the obvious policy conclusion that interventions should take place, as far as possible, in the immediate aftermath of decisions to eschew behaviour that conforms to the simple recommendations made almost universally by health authorities.3 A better known implication of the dissonance formulation for policy has to do with achieving compliance with the rules in ways that do not require outright coercion. The point being that forcing people to comply does not cause a sustainable change of behaviour, because it does not induce the arousal of dissonance itself.4 In many cases, therefore, it is preferable to evoke the desired behaviour change in ways that do not involve forced compliance and do therefore induce a dissonant reaction. This task, however, is no simple matter, as the available evidence all too clearly

2

Aronson and Tavris (2020). See, for example, the preventive measures advocated by the Centers for Disease Control and Prevention (CDC) in the USA and the World Health Organization (WHO) in Geneva. 4 See Sect. 2.4.1. 3

2 Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in. . .

12

Table 2.1 Mask-wearing by country, 2020 Country Germany Canada UK Netherlands Norway USA Australia

Always wear a mask (%) 63 35 19 9 4 59 10

Never wear a mask (%) 6 23 51 60 79 14 64

Source YouGov/Imperial College of London surveys from 22 to 25 June 2020

suggests with regard to the manner in which non-coercive messages to encourage the wearing of masks, social distancing, and so on are widely ignored.5 Table 2.1, for example, shows how often people in a selected sample of (developed) countries say they always wear a mask when leaving home and, conversely, those who say they never wear a mask. Thus, even in a sample of the world’s most developed countries, there is much left to be desired with regard to mask-wearing and (in all likelihood) related preventive behaviour as well.

2.4 2.4.1

Policy Derived from Dissonance Theory Forced Compliance

Under circumstances where the pandemic threatens to overcome the capacity of a country, in terms, say, of medical staff and numbers of beds, there is little alternative (as now in many countries) to forcing compliance with government edicts. Closing down shops, restaurants, and gyms, for example, is a way of helping to enforce compliance with social distancing in areas where this would otherwise be difficult to achieve. ‘Full lockdown’ is an extreme example of this idea which entails confining people to their homes, except for a few exceptional circumstances. The same logic applies to policy which uses incentives to alter behaviour. For, here too, the induced change can be dismissed as an act lacking a volitional component, which, to this extent, would not create any (or very little) dissonance. Note that the higher are the incentives, the lower will tend to be the likelihood that the behavioural change in question will occur. In the commercial sphere, recognition of this logic is known to prevent firms from offering exorbitant incentives to

5

Oddly enough, the Scandinavian countries are the worst offenders in the developed world. ‘According to a recent survey by YouGov, only five to ten percent of respondents in the Nordic countries said they used a mask in public settings’ (Preel and Zakavat 2020).

2.5 Non-Coercive Methods of Compliance

13

purchase their products (Hirschman 1965). In fact, offering no incentive at all may be superior to offering exorbitant ones. In Sect. 2.5, accordingly, I examine methods of compliance that do not involve coercion and which therefore are more likely to arouse dissonance and, as a result, may engender the type of behaviour change that conforms to what the health authorities recommend. Before that, however, it must be stressed that in a situation where so many lives are being lost and non-coercive policies are so widely being ignored, there is much to be said for widespread coercion. For example, countries as diverse as Taiwan, the Republic of Korea, and New Zealand have used coercive policies to great effect in saving lives and reducing the transmission of the COVID-19 virus. In fact, although there were few qualms about massive government intervention during the Second World War, most Western politicians have shown a remarkable reluctance to intervene widely and concertedly during the current pandemic even when the virus is already mutating into new and dangerous forms.

2.5

Non-Coercive Methods of Compliance

Let us return to what was stated earlier as being the core problem of changing behaviour in the dissonance formulation. In particular, ‘While it is necessary to have a minimum of . . . extrinsic motivations in order to obtain compliance . . . the strength of any of these forces beyond the minimum makes a great deal of difference’ (Brehm and Cohen 1962, p. 253). Put otherwise, what policy-makers confront is a vicious circle, or catch-22 situation. That is, that while coercive policies can bring about desired behaviour change, they do not induce the dissonance arousal that is needed for sustained behaviour change. Perhaps, the most appealing way of overcoming this circular problem is through the use of group pressure, which, under certain conditions, ‘may appear as less coercive than pressure exerted directly on individuals from above’ (James and Gutkind 1985, p. 1142, emphasis added).

2.5.1

Group Pressure

These propitious circumstances seem to occur when an individual sees him or herself as belonging to a cohesive group, in whose decisions he participates and in whose norms and values he believes in. ‘Group change will then be easier to effect than change of individuals considered separately’ (James and Gutkind 1985, p. 1124). Unfortunately, however, it seems that in many countries the appeal by the authorities for behaviour change during the pandemic has been pitched at the individual rather than group level. This has, of course, much to do with the distinction between individualistic and more collectivized societies. Table 2.2 is

14

2 Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in. . .

Table 2.2 Top ten countries in terms of masks worn on leaving the house Country (descending order) Philippines Mexico Spain Hong Kong Italy Thailand Japan Malaysia Vietnam Germany

Population always wearing masks (%) 92 85 84 83 83 83 77 77 71 63

Source YouGov/Imperial College of London surveys from 22 to 25 June 2020

instructive in this regard because it shows the prominence of the Asian region in the top ten countries in terms of their proclivity to wear masks consistently. In explaining the comparative success of the (East) Asian countries, An and Tang (2020) have argued in part that, During a pandemic, many policy measures impose extraordinary demands on citizens. For these measures to succeed, public cooperation and voluntary compliance are needed. In addition, citizens need to accept heavy penalties handed out by the government to ensure close to universal compliance . . . Although these actions may look extreme in Westerners’ eyes, they are viable in East Asia due to its collectivist culture. In this culture individuals are willing to sacrifice their freedom during a crisis for the collective good (An and Tang 2020, p. 797).

Historically, both developed and what were then developing countries have used peer pressure in small groups to achieve behaviour change. But the way in which this was achieved demanded a subtle blend of psychological forces. In the successful case of sanitation technology in rural Mexico, for example, the group model used was very similar to that adopted in China. In both cases, that is to say, ‘Peer pressure . . . appears to have been perceived by participants as something much less coercive than a mere directive imposed from above’ (James and Gutkind 1985, p. 1145). Instead, ‘Proposals are discussed thoroughly, and opposing opinions are heard respectfully . . . Never is a vote taken, and as a result, polarization is minimized’ (Elmendorf and Buckles 1980, p. 14, emphasis added). Thus, it was that a concern with solidarity went along with a pattern of social pressure to adhere to collective decisions. Looking back to what occurred during the Ebola crisis in West Africa during 2014 and 2015, van Bavel et al. (2020) point to the role played by trust in securing the desired change in behaviour towards measures aimed at containment of the disease. They draw attention for example to the enlistment of ‘local voices to help build engagement and trust in health officials’ and the way in which this can raise the likelihood of compliance with public health policies (van Bavel et al. 2020, n.p.). As an example they cite a finding that specialized Ebola treatment facilities employing

2.5 Non-Coercive Methods of Compliance

15

‘community liaisons and social mobilizers to raise awareness and resolve misconceptions were associated with increases in reporting Ebola cases’. Other research in Africa—on vaccinations, social distancing, and the use of clinics—attests to the positive role of trust in decisions to behave in a manner that is consistent with government policies during pandemics. In this context, the role of religious groups and leaders perhaps warrants special mention. ‘During the West African Ebola crisis, for example, religious leaders across faiths in Sierra Leone advocated for practices such as handwashing and safe burials. The engagement of the faith-based sector was considered a turning point in the epidemic response’ (van Bavel et al. 2020, n.p.).6 That this sector should have exerted such a major influence (in this case) is not particularly surprising, since religious groups typically play an especially telling role in the socio-economic life of those living in developing countries (and the more so in rural areas of those countries).7 Conversely, and most acutely, a lack of trust in the institutions and leadership in the USA attests to the widespread lack of compliance with the simple preventive measures described above. Indeed, it is difficult to recall a president of that country who has engendered less trust than the former occupant of that office and his administration. Equally, it is nowhere more obvious than in that country that ‘Actions that divide the leader from followers or that suggest that the leader is not prepared to share the burdens of followers can be corrosive to their ability to shape followers’ behaviour’ (van Bavel et al. 2020, n.p.).

2.5.2

The Foot-in-the-Door Technique

A second way of seeking to achieve behaviour change without coercion is known as the foot-in-the-door technique. In fact, the first article on the subject, by Freedman and Fraser (1966), is entitled ‘Compliance without pressure: the foot-in-the-door technique’. The basic idea is very simple: it is that once someone has agreed to a small request, he or she is more likely to accede to a (much) larger one. Freedman and Fraser (1966) ran several experiments to test the theory which was convincingly supported in a number of different formats.

6 Also noteworthy is the attempt by Green et al. (2006) to understand the success in Uganda, of reducing the severity of the HIV/AIDS epidemic in the 1980s. They conclude that that country’s approach to behavioural change relied primarily on community-based and face-to-face communication . . . ‘Strong nongovernment organization (NGO) and community-based support led to flexible, creative and culturally appropriate interventions that helped facilitate individual behavior change as well as changes in community norms’ (Green et al. 2006, n.p.). 7 Religious groups also proved important in the successful experience of Uganda in achieving behaviour change with regard to the use of condoms, during an outbreak of HIV/AIDS there (Allen and Heald 2004).

2 Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in. . .

16

Since then, the technique has been used in a variety of settings, ranging from marketing and politics, to non-profits, with a remarkable degree of success. ‘A political candidate might ask people in attendance at a rally to wear a pin to promote his campaign later, he might ask them for a campaign donation. A group of women may agree to a short health survey, and later agree to breast cancer screening’ (Patel 2014, n.p.). Quite why the theory works in practice, however, has been the subject of debate among those in the field.8 Some see it, for example, as a reflection of the need for cognitive consistency, i.e. that people prefer not to contradict themselves. Freedman and Fraser themselves suggested that ‘What may occur is a change in the person’s feelings about getting involved or about taking action. . . . He may become, in his own eyes, the kind of person who does this sort of thing, who agrees to requests made by strangers, who takes action on things he believes in, who cooperates with good causes’ (Freedman and Fraser 1966, p. 201). To the best of my knowledge, however, no efforts to apply this technique to the COVID-19 pandemic in the Global South have yet been made. The point would be to make requests for mask-wearing, social distancing, and hand-cleaning more appealing than is customary, by first getting from the target groups approval for a smaller request on the topic (such as reading a short statement of the issues). Such a suggestion, I would emphasize, should not be taken lightly, partly because rates of compliance with health requests are still low in many countries in the region and also because there is already some experience with the technique in the aid and commercial sectors.

2.5.3

Induced Hypocrisy

I have already discussed how a clash between one’s beliefs and one’s behaviour gives rise to the unpleasant sensation of dissonance and subsequent efforts to reduce it. But depending on its severity and individual tolerance to it, dissonance can last for quite a long time and in a pandemic-type situation that the world now confronts, the process clearly needs to be abbreviated. This third policy option is also thus rooted in dissonance theory, but an attempt is made to expedite the point at which the individual changes his or her behaviour in the favoured direction. First proposed by Aronson and his colleagues in the early 1990s, the induced hypocrisy idea has been applied to a number of health-related areas such as smoking and condom use in the context of an HIV/AIDS crisis (Aronson et al. 1991). Consider the procedure that was involved in an attempt to invoke induced hypocrisy to promote the use of condoms among young adults (Stone et al. 1994). Thus, Dissonance was created after a proattitudinal advocacy by inducing hypocrisy—having subjects publicly advocate the importance of safe sex and then systematically making the

8

See, for example, the discussion in Scott (1977).

2.5 Non-Coercive Methods of Compliance

17

subjects mindful of their own failures to use condoms. It was predicted that the induction of hypocrisy would motivate subjects to reduce dissonance by purchasing condoms at the completion of the experiment. The results showed that more in the hypocrisy condition bought condoms and also bought more condoms, on average, than subjects in the control conditions (Stone et al. 1994, abstract).

Recently, a literature review of twenty-nine published and nine unpublished induced hypocrisy studies showed that the case just described is far from exceptional in its findings. On the contrary, the review supported the ‘idea that hypocrisy (vs. control) increased both behavioral intention and behavior’ (Priolo et al. 2019, abstract). A weakness in the literature, however, is that it is focused almost entirely on experimental conditions and operational questions have been very largely ignored. Yet, such questions need to be addressed, if only because of the two-stage procedure that is emblematic of the induced hypocrisy paradigm, namely, that ‘the subject is [first] asked to produce a discourse in favour of the social norm being considered and then recall past behaviours that were inappropriate, that is, that did not comply with that norm. According to the induced hypocrisy paradigm, by drawing attention to the contradiction between what people publicly support and what they do, subjects will spontaneously tend to modify their behaviour in a way that makes it more consistent with the norm’ (Odou et al. 2018, n.p.). There are, however, several cases which were conducted in real-life conditions. One of them (on the environment) was carried out at a university swimming pool, and the first part of the induced hypocrisy test took the form of asking participants whether, ‘when they took showers, these were always kept as short as possible, or whether they were sometimes allowed to linger beyond this limit. In the second stage of the test, those who took part, were requested to add their signatures to a flyer advocating water conservation on campus. As predicted, the group that had been subject to induced hypocrisy, significantly reduced the time they spent in the shower’ (Odou et al. 2018, n.p.). In the health sphere as well, research has shown that induced hypocrisy can be made to work in real life, i.e. outside the lab. One case involved 127 members of a college fitness centre, who were requested to supply reasons why they did not exercise regularly, after which they were asked to sign a ‘large poster advocating the regular use of the fitness facility’ (Bator and Bryan 2009, n.p.). ‘The results showed that members in the hypocrisy condition reported significantly higher intentions to exercise regularly, and based on the number of times they swiped their identification card at the entrance, they were more likely to use the recreational facility during the next week, compared to control conditions’ (Bator and Bryan 2009, n.p.). Given its impressive experimental record, on the one hand, and the marked failure of the public to respond positively to informative messages on protection against COVID-19 in most poor countries, on the other hand, one would have thought that the induced hypocrisy paradigm might already have been tried. Yet, as far as I can tell, this has not occurred, even on a modest scale.

18

2 Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in. . .

Perhaps the most obvious reason is that the disease is still relatively new and on this account has not yet allowed research to be organized, conducted, and published. Relatedly, the induced hypocrisy paradigm demands much more contact with subjects during a two-stage sequence than general informative messages on television and the Internet. For this reason, policy based on the paradigm will tend to have a more limited reach than other measures, but its effectiveness may be much higher. There is certainly much that can be learned from existing research conducted in real-life (out of the lab) settings about policy towards health situations and I urge that these be drawn as expeditiously as possible.

2.5.4

The Dynamics of Dissonance Policy

It was noted above that a pernicious dynamic exists within the dissonance formulation after a decision has been made. That is, that there is an inherent tendency for views held at the moment of decision, to become ever more certain as the process endures.9 If the original decision was mistaken, therefore, it becomes increasingly more difficult to reverse. For this reason, I argued that countervailing policy needs to be adopted as soon after the decision as possible. In some countries, however, it is unclear what form such policy should take. The USA, for example, was beset by highly inconsistent public messages and a severe lack of trust in the upper echelons of government (which have gone so far as to decry the very existence of the virus and therefore the need for preventive measures). This diminishes the dissonance that would be forthcoming in a country with a clear, authoritative stance on the pandemic. For its part, the UK has also suffered from these deficiencies and like the USA has a very poor record on societal adherence to generally accepted preventive measures. Many Asian countries, by contrast, with much superior compliance records, have behaved very differently. In Vietnam, for example, ‘The government has taken swift and strong actions against rule-breakers and has provided extensive and accurate reporting to the media so that people are generally well-informed about COVID-19 and the preventative measures they should take. The international community applauded Vietnam, saying that the timely response to the crisis was critical in the early stages (OECD 2020, n.p.). These circumstances would, in turn, have made it difficult for those who were not observing the guidelines to dismiss the virus as minor or non-existent and so to prevent the arousal of dissonance that the theory predicts (under non-coercive conditions).

9

There is some affinity here with Myrdal’s (1957) theory of cumulative causation, according to which initial inequalities in the social system tend, without countervailing policy, to worsen rather than improve over time.

2.6 Conclusions

2.5.5

19

Nudging

I include nudging in the policy list because like the previous alternatives, it is basically a non-coercive method of behaviour change. It is non-coercive because it alters only the way in which messages are presented rather than the range of options themselves. Yet, according to Kim et al. (2020, n.p.), ‘Although people have rapid and constant access to information concerning disruptive events and natural disasters, consumer psychology research and health promotion have not focused on the best way of communicating with individuals in terms of uncertainty.’ The authors themselves, however, have contributed to this knowledge gap by showing that adding relevant comparative data (such as that involving car accidents or flu incidence) leads to a reduction in the perceived threat of the COVID-19 virus (and hence to the common practice of stockbuilding during an epidemic). For the details of this result and a discussion of the role of nudging in relation to the COVID19 pandemic, I refer the reader to the paper mentioned above.

2.6

Conclusions

Cognitive dissonance is probably the most influential theory in social psychology. And although it has been used to understand elements of the COVID-19 pandemic, its policy contributions to the disease have been limited, even as severity, measured in cases and deaths, continues to grow alarmingly in most rich and poor parts of the world. The purpose of this chapter, accordingly, has been to systematically focus on how the dissonance formulation can systematically be used to increase compliance with the requirements of mask-wearing, social distancing, and handwashing. The choice of dissonance theory for this exercise was not based on its prominence, but rather because according to two experts on the topic, the dynamics of the theory are ‘playing out during the pandemic among the many people who refuse to wear masks or practice social distancing’ (Aronson and Tavris 2020, n.p.). Indeed, so strong is the underlying dynamic that when ‘the facts clash with their preexisting conditions, some people would sooner jeopardize their health and everyone else’s than accept new information or admit to being wrong’ (Aronson and Tavris 2020, n. p.). In some cases, the consequences of such behaviour are so severe that governments need to force their citizens to undertake what they (the governments) regard as necessary (amounting on occasion to a ‘full lockdown’). Such cases might arise, for example, when the demands for protective equipment and hospital beds run far ahead of supply. But while ‘forced compliance’ may help to temporarily stabilize the situation, it will not lead to longer-term behavioural change because it can easily be dismissed as ‘something that had to be done’ and thus occasions no dissonance arousal (a necessary condition for sustained behavioural change).

20

2 Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in. . .

Nonetheless, I was at pains to stress that under the current dire conditions, with a mutation of the virus into new and dangerous forms, there is a strong case for widespread coercion on the part of governments. In spite of the successful examples of certain Asian and other more democratic countries, many liberal politicians have shown a disdain for widespread intervention, an attitude that has cost much in the way of lives lost and increased transmissions of the virus. Yet, such politicians seem to have few reservations about imposing widespread coercive policies such as rationing, price controls, conscription, and so on during wartime conditions. On the other hand, though, widespread coercion does not arouse dissonance as argued above, and will not therefore occasion a sustained change of behaviour, and it may also be accompanied by open dissension among the affected population (as recent events in numerous countries illustrate), who are forced to comply with widespread regulation. The rest of the chapter was devoted to non-coercive modes of behaviour alteration. The first of these, the use of group pressure, was adopted extensively in some of the early Asian developing countries, and accounts for much of their superior performance over individualistic cultures in the West. I did stress though that the groups needed to be carefully designed with regard to how the individuals saw their role in them (arising out of discussion, criticism, and cooperation). In individualisticoriented countries on the other hand, far more scope exists than is thought, for groupbased peer pressure, as regards compliance with COVID-19 measures.10 The same goes for another technique of non-coercive compliance, known as the ‘foot-in-the-door’ technique, first postulated in the 1960s (Freedman and Fraser 1966). The basic idea is remarkably simple, namely that an initially small request, if honoured, will facilitate a subsequent larger one. In the health sector, for example, a group of women may first be asked to complete a short health survey and then to subject themselves to screening for breast cancer. The hypothesis is that acceptance of the latter will be more likely to occur because of the agreement to the first request. The foot-in-the-door technique has been widely used in marketing, politics, and fund-raising, but it has not to the best of my knowledge been used with respect to compliance with the COVID-19 regulations. Given the generally limited success of other policies in this regard, it seems worthwhile to try another technique based on dissonance theory as well, even if it is relatively labour-intensive (because it involves two distinct phases). Yet, another possibility is an extreme version of dissonance theory, known as ‘induced hypocrisy’. Essentially, it relies on the fact that people do not always ‘practice what they preach’. The confrontation between what they preach and what they do gives rise to dissonance and hopefully also behaviour change in the desired direction. For example, one could get people to sign a petition in favour of

10 According to van Bavel et al. (2020, n.p.), for example, ‘Perceived norms are also most influential when specific to others with whom common identities are shared, including for the spread of health behaviors. Therefore, messages that provide in-group models for norms (for example, members of your community) may therefore be most effective.’

References

21

mask-wearing, but then point out the occasions when they did not in fact comply with this view in reality. Numerous experiments have confirmed the predicted results of this technique, but I am not aware of any attempts to apply it to the Corona pandemic. Yet, this recommendation too may improve the degree of compliance with existing recommendations. Penultimately, I turn to a dynamic aspect of the dissonance process, which complicates most of the policies I have described above. In particular, it concerns the tendency of beliefs to become more certain over time and hence less amenable to change, if they were initially misguided. The task of policy is therefore to initiate countervailing measures at an early stage of the pandemic, as numerous Asian countries seem to have done with the provision of timely and accurate information about the virus and what should be done about it. Many countries, by contrast, have now to deal with inaccurate beliefs that have only become more rigid over time. The last non-coercive policy that I briefly discussed was nudging; in particular, the way in which the public service message is constructed can have an important influence on the way in which it is perceived and acted upon. There seems, however, to be considerable scope for further research on this technique in developed as well as developing countries. And being concerned with the design of public service messages towards COVID-19, it should not be unduly expensive.

References Allen T, Heald S (2004) HIV/AIDS policy in Africa: what has worked in Uganda and what has failed in Botswana? J Int Dev 16(8):1141–1154 An B, Tang S (2020) Lessons from COVID-19 responses in East Asia: institutional infrastructure and enduring policy instruments. Am Rev Public Adm 50(6–7):790–800 Aronson E, Tavris C (2020) The role of cognitive dissonance in the pandemic. The Atlantic. Available https://www.theatlantic.com/ideas/archive/2020/07/role-cognitive-dissonance-pan demic/614074/ Aronson E, Fried C, Stone J (1991) Overcoming denial and increasing the intention to use condoms through the induction of hypocrisy. Am J Public Health 81:1636–1638 Bator R, Bryan A (2009) Revised hypocrisy manipulation to induce commitment to exercise. Poster presented at the 21st annual conference of the american psychological association, San Francisco, CA, May 22–25 Brehm J, Cohen AR (1962) Explorations in cognitive dissonance. Wiley Cooper J (2019) Cognitive dissonance: where we’ve been and where we’re going. Int Rev Soc Psychol 32(1):1–16 Elmendorf M, Buckles P (1980) Appropriate technology for water supply and sanitation: sociocultural aspects of water supply and excreta disposal. Available https://agris.fao.org/agris-search/ search.do?recordID¼US2012400020 Freedman JL, Fraser SC (1966) Compliance without pressure: the foot-in-the-door technique. J Pers Soc Psychol 4(2):195–202 Green EC, Halperi DT, Nantulya V, Hogle JA (2006) Uganda’s HIV prevention success: the role of sexual behavior change and the national response. AIDS Behav 10:335–346 Hirschman A (1965) Obstacles to development: a classification and a quasi-vanishing act. Econ Dev Cult Change 13(4):385–393

22

2 Cognitive Dissonance Theory and Policy Towards the Corona Epidemic in. . .

James J, Gutkind E (1985) Attitude change revisited: cognitive dissonance theory and development policy. World Dev 13(10–11):1139–1149 Kim J, Giroux M, Gonzalez-Jimenez H, Jang S, Kim S, Park J (2020) Nudging to reduce the perceived threat of coronavirus and stockpiling intention. J Advert 49(5):633–647 Myrdal G (1957) Economic theory and underdeveloped regions. Gerald Duckworth, London Odou P, Darke P, Voisin D (2018) Promoting pro-environmental behaviours through induced hypocrisy. Rech Appl Mark (English edition) 34(1):74–90 OECD (2020) COVID-19 crisis response in ASEAN states. Available https://oecd.org/coronavirus/ policy-responses/covid-19-crisis-response-in-asean-member-states-02f828a2/ Patel N (2014) Foot-in-the-door technique: how to get people to seamlessly take action. Forbes. Available https://www.forbes.com/sites/neilpatel/2014/10/13/foot-in-the-door-technique-howto-get-people-to-take-seamlessly-take-action/?sh¼484129e37d9e Preel M, Zakavat, N (2020) Northern exposure: Nordic countries defy rest of world to go mask-free. The Times of Israel. Available http://www.timesofisrael.com/northern-exposure-nordiccountries-defy-rest-of-the-world-to-go-face-mask-free/ Priolo D, Pelt A, St. Bauzel R et al (2019) Three decades of research on induced hypocrisy: a metaanalysis. Personal Soc Psychol Bull 45(12):1681–1701 Scott CA (1977) Modifying socially conscious behavior: the foot-in-the-door technique. J Consum Res 4(3):156–164 Smith LGE, Gibson S (2020) Social psychological theory and research on the novel coronavirus (COVID-19) pandemic: introduction to the rapid response special section. Br J Soc Psychol 59 (3):571–583 Stone J, Aronson E, Crain AL, Winslow MP, Fried CB (1994) Inducing hypocrisy as a means of encouraging young adults to use condoms. Personal Soc Psychol Bull 20(1):116–128 van Bavel JJ, Baicker K, Boggio PS et al (2020) Using social and behavioural science to support COVID-19 pandemic response. Nat Hum Behav 4:460–471 YouGov (2020) Imperial College of London, surveys from June 22–25

Chapter 3

Digital Divide Reversal: Evidence, Explanations, and Implications

3.1

Introduction

The global digital divide customarily refers to the differential extent to which rich and poor countries have access to digital technologies and the complementary resources that are required to make effective use of them. In fact, access is a necessary condition for benefits to be derived from these technologies, and it is not surprising that the penetration of the Internet is skewed in favour of the rich rather than the poor countries, as shown in Table 3.1. After all, such technology is, in large part, developed in and for conditions in the former category. (This is known as the first digital divide and it concerns access rather than use of technology.) Given the linear relationship between income and Internet use, it is the poorest countries in which the comparison with the richest group is most marked (16% as against 87%). For measurement of the second digital divide, I have selected digital skills as the most relevant variable, given that it is often described as the main limitation to extracting the benefits of the Internet, once it is adopted (van Dijk and van Deursen 2014). Table 3.2 shows the extent of this divide with reference to basic and standard skills, as defined by the ITU. In this case, too, therefore, there is a distinct divide, especially with respect to standard skills. But whereas there has been a great deal of attention paid to both the first and second divides, little has been written about the possibility of a case where it is the poor countries that appear to derive more from the Internet than the rich: what I refer to as a digital divide ‘reversal’.1 Specifically, this occurs in a situation where the former countries spend more time on the Internet than the latter. Indeed, the purpose

1

It is sometimes mentioned (e.g. by Statista) but is rarely analysed as far as I can tell. Such analysis is the novel feature of this chapter. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. James, New Perspectives on Current Development Policy, SpringerBriefs in Economics, https://doi.org/10.1007/978-3-030-88497-0_3

23

24

3 Digital Divide Reversal: Evidence, Explanations, and Implications

Table 3.1 Internet use by region (% of population, 2017/18)

Region High-income Upper middle Lower middle Low

Year 2018 2017 2017 2017

Internet use (%) 87 56 42 16

Source World Bank, data (CC-BY 4.0) Table 3.2 The digital divide in skills, 2017 Type of digital divide Basic Standard

Developed countries (average proportion of individuals in %) 65 49

Developing countries (average proportion of individuals in %) 46 20

Source ITU (2018), Measuring the Information Society report, Chart 2.2

of this chapter is to provide evidence on the phenomenon, to explain whether and why it occurs, and to discuss the main implications of the findings.

3.2

Digital Divide Reversal: The Data

There are abundant data on the number of people in a country who spend more than a given minimum of time on the Internet (e.g. PEW 2020). But while such a measure may be useful for some types of cross-country analysis, it does not suit my purposes. More specifically, while I am concerned with the actual amount of time people spend on the Internet in rich and poor countries, the measure just described provides information on only the number of users of the technology. As such, it is more akin to an access than a use measure. The type of data that are required to measure the difference in time spent on the Internet are relatively scarce, but they do exist. Table 3.3, for example, shows a list of countries ordered by time spent on the Internet above and below the worldwide average. It also provides the per capita incomes of each country reported. If one can properly speak of a digital divide reversal, the average of countries above the world average would need to have average incomes below those in the group of lower than average countries. Effectively, therefore, this exercise can be regarded as a test for the existence of digital divide reversal. Thus, in terms of time spent on the Internet per day, the Philippines heads the list with just over 10 h, while at the other extreme, Japan records an amount of only 3.45 h. On the other hand, whereas income per capita of the former country is only US$3.485, it is over US$40,000 in the high-income case. The size of this disparity hints at a digital divide reversal, but of course one needs to consider the results for all countries shown in Table 3.3. These are reported in Table 3.4. Thus, although countries below the average have incomes that are more than three times higher than those in the other group, they spend almost 3 h less per day on the

3.2 Digital Divide Reversal: The Data

25

Table 3.3 Time spent per day on the Internet and income per head (2019) Country (in descending order of time spent on the Internet) Philippines Brazil Thailand Colombia Indonesia South Africa Argentina Malaysia Mexico UAE Egypt India Turkey Singapore Saudi Arabia Vietnam Worldwide Portugal USA Russia Hong Kong Italy Poland Sweden New Zealand Ireland China Canada UK Denmark Spain South Korea Australia Austria Belgium Switzerland Netherlands France Germany Japan

Time spent on the Internet per day (in hours) 10.02 9.29 9.11 9.00 8.36 8.25 8.19 8.05 8.01 7.54 7.53 7.47 7.15 7.02 6.44 6.42 6.42 6.38 6.31 6.29 6.23 6.04 6.02 5.56 5.55 5.54 5.52 5.51 5.46 5.28 5.18 5.14 5.04 5.01 5.01 4.58 4.44 4.38 4.37 3.45

Source World Bank data (2020), license CC-BY 4.0 Note Income per head in current US$

Per capita income in US$ 3485.1 8717.2 7808.2 6432.4 4135.6 6001.4 10,006.1 11,414.8 9863.1 43,103.3 3020.0 2104.1 9042.5 65,233.3 23,139.8 2715.3 23,145.0 65,118.4 11,585.0 48,755.8 33,189.6 15,595.2 51,610.2 42,084.4 78,661.0 10,261.7 46,194.7 42,300.3 59,822.1 29,613.7 31,762,0 54,907.1 50,277.3 46,116.7 81,993.7 52,447.8 40,493.9 46,258.9 40,246.9

26

3 Digital Divide Reversal: Evidence, Explanations, and Implications

Table 3.4 Aggregate results for countries above and below the worldwide average, 2019 (for time on the Internet) Time spent on the Internet per day Above average Below average Difference

8h 5.3 h 2.7 h

Income per capita 13,513.9 43,584.4 30,070.5

Source World Bank data (2020), license CC-BY 4.0 Table 3.5 Time spent on social media and the Internet as a whole Country Philippines Nigeria India China USA UK Germany Japan

Time spent on social media (per day in hours) 4.01 3.36 2.25 2.19 1.57 1.5 1.15 0.45

Total time spent on the Internet (per day in hours) 10.02 n/a 7.47 5.52 6.31 5.46 4.37 3.45

First column as % of second column 40% 30% 40% 25% 27% 26% 13%

Source World Economic Forum, Global Web Index (2019); World Bank data (2019), license CC-BY 4.0

Internet. Bearing in mind the caveat that the sample of countries is limited, the results point clearly to what I described earlier as a digital divide reversal. Although, as also noted above, it has attracted very little attention in the literature, the topic has not gone entirely unnoticed. Hughes (2019, n.p.), for example, finds it curious that ‘when the figures are broken down by country, there’s an interesting trend that shows the most enthusiastic Internet users being located primarily in developing and middle-income countries’. He goes on to note, as we have, that ‘Internet users in the Philippines, for example, spend 10.02 h online each day, with Brazil trailing behind at 9.20 h. In comparison, Japanese users are connected for just 3.45 h, while French people spend just 4.38 h online’. In relation, specifically to social media usage on the Internet, another observer has noted that ‘You might think that the richer a country is, the more likely it is to have people using social media. But this isn’t necessarily the case’ (Digital Marketing Institute 2020). In fact, when one looks at the amount of time spent on social media by different countries, it is low- or low-middle-income countries that occupy the first three places, as shown in Table 3.5. The table also indicates the amount of time spent on this activity as a percentage of total time on the Internet. Thus, in the countries which spend the most time on social media activities, this amount can run to as much as 40% of the total number of hours devoted to the Internet. More generally, the evidence points to the rather paradoxical conclusion that developing countries are given to spending their Internet time on entertainment,

3.3 Explanations Table 3.6 Distribution of video gamers worldwide, by age group, 2018

27 Age range 18–20 21–35 36–50 51–65

Gamers (%) 9 39 33 19

Source Statista (2020)

rather than developmental activities such as searching for information on health, government services, and vacancies for jobs.2 A recent finding for Nigeria is that new users of the Internet ‘tend to use mobile internet for entertainment. Career development, personal health management, and work are little known uses’ (Leung 2020, n.p.). I turn next to examine some possible reasons for the digital divide reversal in Internet use described above. Note that if we are to explain this phenomenon, it will require a focus on those uses of the technology in which poor and low-middleincome countries spend more time on the Internet, rather than those in which the opposite is true. (That is, where more time is spent by the developed countries.)

3.3 3.3.1

Explanations Demographics and the Popularity of Gaming

As far as the gaming component of social media use is concerned, the argument hinges on two related propositions. One is that much of the activity is undertaken by the more youthful members of society and the other is that a pervasive feature of poor countries is precisely the relative abundance of such persons in the age structure. Let us then begin by providing evidence in support of the first proposition. Table 3.6 shows the worldwide distribution of video gamers by age group in 2018. Thus, the age range 18–35 years comprises 48% of total users worldwide and those in the 21–35 group make up the largest category of users. That developing countries tend to have a low median age compared to richer ones can be illustrated with reference to the selected country sample shown in Table 3.7. Although the samples are admittedly very small, the difference between the average for developed and developing countries is certainly stark. Indeed, the median age for the latter is almost one and a half times as low as that for the former group. Taken together, Tables 3.6 and 3.7 suggest that there will be a more extensive

2

One should not be too hasty in condemning the choice of entertainment by the poor. Douglas and Isherwood (1982), for example, argue that goods are a means of communicating with one another. Banerjee and Duflo (2006) point to the surprising amount spent on festivals by the poor in an Indian village.

28

3 Digital Divide Reversal: Evidence, Explanations, and Implications

Table 3.7 Median age by country (selected sample) in 2017

Median age Developed countries Japan Germany Italy Austria Spain Average Developing countries Costa Rica Vietnam India South Africa Bangladesh Average

48.6 47.8 46.5 44.5 43.9 46.2 32.6 31.9 28.7 28.0 27.9 29.8

Source CIA (2018) Note Developing countries also include those with low-middle incomes

use of the gaming component of social media in low- and low-middle-income countries, than in those that are relatively affluent. There is thus here also a partial explanation of the digital divide reversal referred to above.3

3.3.2

The Differential Value of Time

This explanation is based on the recognitions that entertainment is a relatively timeintensive activity and that time is a comparatively abundant resource in developing countries, especially among those with the lowest income levels. Becker (1965), for example, assumes that the ‘value of an hour equals average hourly earnings’ and thus that there is a one-to-one correspondence between earnings and the value of time. ‘Time is thus much more valuable in rich as compared to poor countries’ (James 2014, p. 71). The opportunity cost of leisure thus tends to be lower in countries with relatively low wages and high rates of unemployment and under-employment. That is, that what is forgone in the pursuit of leisure and entertainment in low-income countries is generally not high-paying jobs, but rather those with low wages in the informal sector of the economy.4 In fact, some see the pursuit of entertainment on the Internet as a pleasurable means of passing time in the so-called timepass economy, among

3

The younger age structure in developing countries is caused by higher fertility rates, compared to developed countries. 4 Forgone opportunities also include open unemployment or unemployment that is disguised.

3.3 Explanations

29

Table 3.8 Quality of overall infrastructure index (2017) (7 is best) Developed countries Australia Austria Germany Italy Norway Portugal Switzerland UK USA Belgium Canada Finland Spain France Average

4.7 5.9 5.7 4.3 5.2 5.7 6.7 5.0 5.9 4.9 5.2 6.1 5.5 6.0 5.5

Developing countries (low and low-middle income) Bangladesh 2.9 Benin 2.4 El Salvador 3.3 India 4.6 DRC 2.0 Cameroon 2.3 Nepal 2.9 Pakistan 3.8 Philippines 3.0 Sierra Leone 2.6 Sri Lanka 3.9 Tanzania 3.6 Vietnam 3.6 Republic of Yemen 2.2 Average 3.1

Source World Economic Forum (2018), Global Competitiveness Index

those with limited digital skills and an abundance of time on their hands (where timepass is the Indian-English word for killing time). According to The Economist (2019), for example, ‘For many people the phone offers an unsurpassable opportunity for turning otherwise empty time into something enjoyable. . . . the internet is the leisure economy of the world’s poor.’ And as we have already noted, much of the period spent on passing time is devoted to entertainment in general and video games in particular. In fact, ‘the vast majority of the top 25 apps by revenue in both Google’s and Apple’s app stores are games . . . Tencent became one of China’s internet giants because of games. Facebook grew into the world’s sixth-most valuable company by giving people a place to “do timepass” . . . the fastest-growing new apps of recent years have all been aimed at timepass’ (The Economist 2019).

3.3.3

Communications and Infrastructural Differences

Under this heading, I advance the proposition that the digital divide reversal has much to do with the relative lack of infrastructure in poor and medium-income countries, as it relates to the communication opportunities afforded by the Internet. Because of a lack of country comparative data on specific components of the infrastructure, however, I rely instead on the equality of overall infrastructure index, compiled by the World Economic Forum for its Global Competitiveness Index. On this basis, the scores for a sample of low- and middle-income countries are compared with those of a group described as developed. The results are shown in Table 3.8.

30

3 Digital Divide Reversal: Evidence, Explanations, and Implications

Predictably, the average score for the developed countries is sharply higher than that of the group with lower incomes. In fact, this finding holds true not just for the low-income countries such as the Democratic Republic of Congo (DRC) and Cameroon, but also for the low-middle-income groups such as Nepal and Tanzania. Assuming, as one surely can, that this aggregate result applies also to transport infrastructure such as buses, trains, taxis, and planes, then implications begin to emerge for the differential extent to which the Internet is used in rich and poor countries. The main one from my point of view was in fact apparent with basic mobile phones well before the advent of the Internet. As described at the time the hypothesis was that, the more difficult and expensive it is to communicate by means other than [basic] mobile phones, the more (percentage) use will be made (ceteris paribus) of this technology. And since it is assumed that the said difficulty varies inversely with country income, the prediction is that the most intensive use of mobile phones will occur in relatively poor countries (James 2014).

This hypothesis has been tested for three different developing countries, with varying levels of income and transport infrastructure (Samuel et al. 2005). The countries are Tanzania, South Africa, and Egypt. The authors draw attention to the travel time and costs that are saved by a mobile phone, especially in countries where the opportunities for public modes of transport are particularly limited. Thus, in such situations, the savings from a mobile phone call will tend to be especially marked. This was found to be true in the survey by Samuel et al. (2005) with respect to a comparison between South Africa and Tanzania, in that the savings in the latter country were higher than in the former, since ‘roads are worse and public transport less extensive’ (Samuel et al. 2005, p. 49). The authors conclude that ‘The potential importance of mobile as a substitute for travel is easy to underestimate. Of the communities surveyed in South Africa, only four out of ten had a regular bus to the nearest town and the typical round-trip cost was 15 Rand. In contrast, a typical pre-paid voice call cost R5’ (Samuel et al. 2005, p. 49). The same basic principles also apply to mobile phones with Internet connectivity and especially to the recent introduction of low-cost smart feature phones which were first introduced in India, where they have subsequently been adopted by tens of millions of inhabitants, who might otherwise have been unable to afford expensive smartphones designed for the rich countries (though the price of these has fallen somewhat in recent years5). In fact, phones with Internet connectivity increase the appeal of communication at a distance, rather than visiting friends and family, because of video opportunities afforded by Zoom and other applications (though, as Gina Porter (2016) has rightly emphasized, there are cultural limits to the extent to which remote communications can substitute for physical presence).6 This is true even where the distance to be travelled is substantial.

5

See The Economic Times|Tech (2020). The point here is that there are some occasions, such as festivals and funerals, where physical presence is required and digital communications are not regarded as a substitute therefore.

6

3.3 Explanations

3.3.4

31

Costs of Internet Data and Affordability

The cost of handsets and Internet data, in relation to the income of potential users, is an important determinant of differential Internet use across countries and as such may help to explain the digital divide reversal, as this was presented above. It is well to emphasize in this regard that the relevant concept here is affordability, not just phone or data costs. For example, even a country facing relatively high costs may enjoy incomes that are substantial enough to enable a wide segment of the population to afford Internet use. The JioPhone in India, however, is an excellent example of the opposite case where costs were so extraordinarily low7 that Internet access was made available even to large numbers of those with low incomes in the rural areas (where the poor often tend to be concentrated). Referred to as a smart feature phone, the Jio is a hybrid between feature phones without Internet connectivity and smart phones, which are designed principally in and for the rich countries.8 As a major example of what can be achieved on a large scale by appropriate technology, the JioPhone was highly successful in India, soon after its introduction there.9 The year 2018, for instance, saw a growth in the demand for smart feature phones of 252% (Counterpoint Research 2019). Furthermore, estimates by Counterpoint Research, a consultancy, predict the sale of almost 370 million smart feature phones worldwide between 2019 and 2021 (Leung 2019, n.p.). Without doubt, the marked success of the JioPhone would not have been possible without the heavy subsidization by Reliance Industries, one of India’s largest conglomerates and one of the owners of the Jio project. These subsidies allowed costs and prices to be much lower than they otherwise would have been. For example, although the handset was priced at Rs 1500, it was effectively free since buyers were reimbursed after 3 years (Financial Express 2017). Apart from the somewhat special Indian case, it is to the Asian region more generally that one most needs to look in order to explain why low- and middleincome countries feature so prominently in the group that spends most time on the Internet. Recall from Table 3.3 that the Philippines spends on average the most time per day on the Internet with Thailand in third place, closely followed by another Asian country, Indonesia. Thus, out of the top five countries, no fewer than three are from the Asian region. Moreover, India and Vietnam also form part of the group that is above average. At the level of the region as a whole, unusually low costs appear to account for much of our understanding of the digital divide reversal. In fact, a survey conducted

7

In fact, Internet costs in India at one point were the lowest in the world (Statista 2019). This means that they embody sophisticated features that are unlikely to be used by those in poor countries where digital skills tend to be sorely lacking. For this and other reasons, they may reasonably be described as inappropriate technology. 9 In explaining the success of the JioPhone one cannot ignore the role of the software developer, KaiOS Technologies, which relied heavily on open-source technology (James 2020). 8

32

3 Digital Divide Reversal: Evidence, Explanations, and Implications

by the Alliance for Affordable Internet in 2019 showed that ‘Asia is the only region that has reached the UN’s ‘1 for 20 threshold for internet affordability—defined as 1GB for no more than 2% of average monthly income. This is the level at which access becomes affordable for most people, including those at below average income levels’ (Alliance for Affordable Internet 2019, n.p.). Of course, the 2% target could also be achieved with higher costs, provided that incomes are high enough, but then the countries concerned would tend to fall into the category of high, rather than the group of low, or low-middle-income nations.

3.4

Conclusions and Implications

The digital divide refers in this chapter to the differential extent to which countries benefit from the Internet. Until now, it is the rich countries that have gained the most, partly because they can afford greater access to this technology and partly because they have more of the complementary resources—such as digital skills—that are required to derive more of the benefits that access promises. The purpose of this chapter, however, has been to raise and analyse an aspect of Internet use that does not conform to this pattern, one which I refer to as a digital divide reversal. Section 3.2 was devoted to investigating the existence and extent of the reversal, which involves the average time spent on the Internet by rich as opposed to low-middle and low-income countries. Contrary to what one might have expected,10 the two latter groups of countries devoted more time on average to using this technology than the more affluent sample. The result presents something of a paradox in that, with the exception of a limited number of commodities, higher incomes are generally expected, in basic economics, to induce greater consumption and welfare. Numerous arguments were then offered as to why the observed reversal occurred. I suggested first that the relative popularity of gaming in poor countries had to do with demographics. On the one hand, that is to say, such countries are known to have a relatively young age structure, while on the other hand, games tend to be played most intensively by precisely the more youthful groups. Other explanations relied on the differential value of time in rich and poor countries and the low opportunity cost of leisure in the latter; the generally poor state of transport infrastructure in the latter, which increases the demand for Internet communications, in place of travel to distant friends and family; and the use of online entertainment to pass time in countries where this resource (time) is in abundance. The welfare implications of these findings are more complex than one might at first suppose. Certainly, the simple assumption that ‘more is better’ cannot be widely justified. Consider, for example, the finding that video gaming is especially popular

10 The expectation is related to the notion of a ‘normal’ good, in economics, i.e. a good for which demand increases as incomes rise.

3.4 Conclusions and Implications

33

in low-income countries, notably, as noted, among the younger members of the population. While some of it may improve welfare, one also has to reckon, however, with the likelihood that some part of this behaviour leads to unhappiness and despair, rather than a feeling of well-being. What I am referring to is the case where gaming becomes compulsive and very difficult to stop. In fact, such a condition has recently been recognized as a mental health disorder by the World Health Organization (WHO) and the Diagnostic and Statistical Manual (the ‘DSM’) of mental disorders, compiled by the American Psychiatric Association (APA). The former defines ‘gaming disorder . . . as a pattern of gaming behavior . . . characterized by impaired control over gaming, increasing priority given to gaming over other activities to the extent that gaming takes precedence over other interests and daily activities, and continuation or escalation of gaming despite the occurrence of negative consequences’ (WHO 2018, p. 1). Estimates of the prevalence of this disorder in developing countries vary because of methodological and other differences, but the problem seems to be especially severe in Asia. Then too, while messaging on platforms such as Facebook may often be a valuable means of enhancing one’s social life, there are also well-known problems of disinformation, hate speech, and various forms of Internet crime. Here, as elsewhere, the goal is to encourage users to make informed choices, but the actors involved—such as government institutions—generally engage in far too little regulation to bring this about.11 If these problems are more acute in developing as opposed to developed countries, the existence of a digital divide reversal may not also lead to a welfare reversal between users of the Internet in the two regions. This is an issue that clearly warrants further research.12 What is clear though is that most governments have paid much more attention to gaining Internet access, than to raising digital skills levels. The result, according to the IFC, is that ‘in Sub-Saharan Africa a significant gap exists across all levels of digital skills in the region, with a lower availability of skills than in other markets . . . The supply of digitally skilled labor in Sub-Saharan Africa . . . must increase to meet anticipated labor market needs or Africa’s economies will falter’ (IFC 2019, p. 5). Redressing this imbalance between supply and demand will require a variety of different policies, involving public and private institutions, which are dependent on the required skills. Some firms have already begun to recruit skilled foreign labour,

There are, however, some notable exceptions to this view. One of them called ‘Life’ was introduced on the Internet by KaiOS Technologies, a software developer. Intended to help inexperienced users, the program ‘offers . . . a directory of curated content such as on woman’s empowerment, health, education, and agriculture’ (James 2020, n.p.). 12 A start in this direction has in fact been made by PEW in its surveys of how users feel about the effects of the Internet on themselves and society as a whole. See, for example, PEW (2020). Much of this research, however, is undertaken with reference to rich rather than poor countries. There are, however, a few surveys of the latter, one of which was published in 2019 (PEW 2019). It showed that although the majority of respondents felt more informed about current events, roughly the same percentage held the view that they were easier to manipulate with false information and rumours. 11

34

3 Digital Divide Reversal: Evidence, Explanations, and Implications

but this is a policy that is neither durable nor beneficial to the local capabilities on which the future will depend.

References Alliance for Affordable Internet (2019) Redefining affordability to achieve universal internet access. Available https://a4ai.org/affordable-internet-is-1-for-2 Banerjee A, Duflo E (2006) The economic lives of the poor. Available http://economics.mit.edu/ files/530 Becker G (1965) A theory of the allocation of time. Econ J 75(299):493–517 CIA World Factbook (2018). Available https://www.cia.gov/library/publications/download/ download-2018/index.html Counterpoint Research (2019) Smart feature phones to create US$28 billion revenue opportunity. Available https://www.counterpointresearch.com/smart-feature-phones-create-us28-billion-rev enue-opportunity/#:~:text¼According%20to%20Neil%20Shah%2C%20Research,over%20the %20next%20three%20years Digital Marketing Institute (2020) Social media: what countries use it most and what are they using? Available https://digitalmarketinginstitute.com/blog/social-media-what-countries-use-it-mostand-what-are-they-using Douglas M, Isherwood B (1982) The world of goods. WW Norton Financial Express (2017) Reliance Jio phone priced free, effectively Rs zero. Available https:// www.financialexpress.com/industry/reliance-jio-phone-price-rjio-phone-free-rs-zero-rs-1500refundable-ril-mukesh-ambani-jio-phone-launch-top-10-specs-and-features-made-in-india/ 772895/ Global Web Index (2019) Social media user trends report. Available https://www.globalwebindex. com/reports/social-2019 Hughes M (2019) Study shows we’re spending an insane amount of time online. Available https:// thenextweb.com/tech/2019/01/31/study-shows-were-spending-an-insane-amount-of-timeonline/ International Finance Corporation (2019) Digital skills in Sub-Saharan Africa. Available https:// www.ifc.org/wps/wcm/connect/industry_ext_content/ifc_external_corporate_site/education/ publications/digital+skills+in+sub-saharan+africa International Telecommunication Union (ITU) (2018) Measuring the information society report. Geneva James J (2014) Internet use, welfare, and well-being: evidence from Africa. Soc Sci Comput Rev 32 (6):715–727 James J (2020) The smart feature phone revolution in developing countries: bringing the Internet to the bottom of the pyramid. Inf Soc 36(4):226–235 Leung E (2019) The birth of the smart feature phone revolution. Available https://www.kaiostech. com/the-birth-of-the-smart-feature-phone-revolution/ Leung E (2020) KaioOS releases findings of research on internet use and perception in Nigeria. Available https://www.kaiostech.com/kaios-releases-findings-of-research-on-internet-use-andperception-in-nigeria/ PEW (2019) Misinformation and fears about its impact are pervasive in 11 emerging economies. Available https://www.pewresearch.org/fact-tank/2019/05/13/misinformation-and-fears-aboutits-impact-are-pervasive-in-11-emerging-economies/ PEW (2020) 8 Charts on internet use around the world as countries grapple with COVID-19. Available https://www.pewresearch.org/fact-tank/2020/04/02/8-charts-on-internet-use-aroundthe-world-as-countries-grapple-with-covid-19/

References

35

Porter G (2016) Mobilities in rural Africa: new connections, new challenges. Ann Assoc Am Geogr 106(2):434–441 Samuel J, Shah N, Hadingham W (2005) Mobile communications in South Africa, Tanzania and Egypt: results from community and business surveys. Vodafone Policy Paper Series 2:44–52 Statista (2019) The cost of mobile Internet around the world. Available https://www.statista.com/ chart/17247/the-average-cost-of-mobile-data-in-selected-countries/ Statista (2020) Share of mobile gamers worldwide as of December 2018. Available https://www. statista.com/statistics/246577/age-distribution-of-mobile-gamers/ The Economic Times|Tech (2020) Smartphone price drops you should know about. Available https://economictimes.indiatimes.com/tech/hardware/smartphone-price-drops-you-shouldknow-about/changing-markets/slideshow/76850356.cms The Economist (2019) A global timepass economy, print edition, June 8 van Dijk JAGM, van Deursen AJAM (2014) Digital skills: unlocking the information society. Palgrave Macmillan World Bank data (2019) License CC-BY-4.0 World Bank data (2020) Creative commons attribution, license CC-BY 4.0 World Economic Forum (2018) Global competitiveness report. Available https://www.weforum. org/reports/the-global-competitveness-report-2018 World Health Organization (WHO) (2018) Gaming disorder. Available https://www.who.int/newsroom/q-a-detail/addictive-behaviours-gaming-disorder

Chapter 4

Why Is India So Dominant in the Demand for New Smart Feature Phones That Are Internet Connected?

4.1

Introduction

For inhabitants of low- and low-middle-income countries, there was, until 2017, only one way of gaining access to the Internet, namely, through smartphones designed mostly for the developed countries. That year, however, saw a major change in the way Internet could be accessed in the form of a so-called smart feature phone, which was basically a hybrid between basic feature phones (without Internet connectivity) and the smartphone, with its complex and expensive characteristics. What had emerged, that is to say, was a classic example of what Schumacher (1973) referred to as an intermediate technology, one which allowed those with lower incomes to gain access to basic Internet features at an affordable price. Indeed, the form of the so-called JioPhone in India was the unusual product of a desire on the part of both major players—KaiOS, an open-source software designer, and Reliance Industries, a major conglomerate, to provide low-cost Internet access to the poor in developing countries. It was unusual in that large conglomerates are not usually given to participate in appropriate technologies on a vast scale. After its introduction in India during 2017, the Jio was an almost immediate success: ‘In 2018, global feature phone demand grew 253% year on year with India being the biggest contributor to this demand’ (Livemint 2019). It is true that the new technology has been introduced in other markets (such as Indonesia, Rwanda, and Tanzania), but so far as I can tell from the limited data, on nothing like the scale achieved in India (even in per capita terms).1 One should further note that there are other very large middle-income countries, where a smart feature phone could have been introduced on a large scale, such as Brazil, China, and Indonesia.

1

KaiOS-based smart feature phones have recently been widely introduced in Africa, but the extent of their success there is not yet widely documented, because too short a time has elapsed in order to do so. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. James, New Perspectives on Current Development Policy, SpringerBriefs in Economics, https://doi.org/10.1007/978-3-030-88497-0_4

37

38

4 Why Is India So Dominant in the Demand for New Smart Feature Phones That. . . Stage / Variables

Generation

Pre-adoption (binary) stop or go

Adoption

Use

Welfare

As noted in the text, both Reliance Industries and KaiOS Technologies were determined to design a product that would cater to prospective users with relatively low incomes in developing countries. This meant retaining basic features of smartphones but eliminating costly “frills”

4G coverage or not (determines whether process continues or stops for those excluded)

Costs of handset and data

Digital skills and literacy

Entertainment vs. development uses

Awareness or not of the mobile Internet on the part of those who are covered (determines whether process continues or stops for those covered but lacking awareness)

Local content

Literacy and digital skills

Relevance

Information on use of the Internet Time spent on different Internet options Voice vs. text for illiterates and elderly

Informed preferences Time spent on Internet Availability of preferred options Local content

Fig. 4.1 The analytical framework

The purpose of this chapter, accordingly, is to explain why India has so dominated the market for KaiOS-based smart feature phones. And towards this end, I employ two related analytical tools, one being an extensive multi-sequential framework and the other being the use of data at each stage (mainly) from the latest issue of the annual GSMA report (2020).

4.2

Methodology

The foundation of the proposed method is a multi-sequential framework in which variables at one stage of the process may influence the outcome at later stages (see Fig. 4.1 below).2 This is perhaps most obviously the case, where binary decisions at an early phase of the process determine whether the project goes ahead or stops completely (as when, for example, there is no awareness of the Internet even though there is complete 4G coverage available). Or again, where variables that influence

2

Figure 4.1 is a much adapted version of a simple figure in James (2018).

4.3 Explaining the Indian Experience with the JioPhone in Comparative Perspective

39

adoption also bear on the welfare effects of the technology (literacy and digital skills come most readily to mind in this context). And in yet another case, the generation of the technology affects not only its adoption but also its impact on welfare. The second element of the methodology is, where possible, to fill in (comparative) data at each stage of the sequential framework shown in Fig. 4.1, which will hopefully help to explain the dominance of India with respect to the smart feature phone. Much use is made in this exercise, of data drawn from the most recent annual survey of mobile phones by the GSMA (2020), which is the most authoritative source on all matters pertaining to these phones. This recent issue of the publication, moreover, pays particular attention to the rapid growth of the smart feature phone in India in the past few years. While this framework draws heavily on economic factors such as costs, it also pays attention to variables from other disciplines such as anthropology, culture, and linguistics. For, it is already known from the development literature on the diffusion of innovations in developing countries that these variables simply cannot be ignored in any attempt to explain their success or failure.3 Even a cursory view of the contents of Fig. 4.1 suggests that it is more complex than the simple generation, adoption, and impact phases. For one thing, it includes a pre-adoption phase with thresholds that determine whether adoption will take place at all. This is apparent with regard to the column headed pre-adoption, in which further progress can be curtailed with reference to a lack of 4G coverage and whether, even with complete coverage, there is awareness of the Internet (GSMA 2020). On these grounds, it may reasonably be argued that policy should focus on these binding constraints, where they are applicable.

4.3 4.3.1

Explaining the Indian Experience with the JioPhone in Comparative Perspective Generation

The generation of innovations still typically takes place in and for the rich countries, where incomes are relatively high and able to absorb products and processes that would usually be considered too expensive, even among those with average incomes in the poorest countries in the developing world (GSMA 2020). The result is, of course, that developing countries are, to a greater or lesser extent, excluded from the benefits of global innovation (apart, that is, from the richest groups in some of these countries). It bears noting that the smartphone developed mostly in rich countries contains many features that are suitable for those countries but not those that are much less well developed (especially those parts of them that are themselves especially poor 3

See, for example, Zanello et al. (2015).

40

4 Why Is India So Dominant in the Demand for New Smart Feature Phones That. . .

and marginalized). In the form, say, of Apple’s iPhone they are affordable to only a small minority of those living in relatively backward countries. Prior to the invention of the JioPhone, therefore, most of the world’s citizens were left totally without access to the Internet (and note that there are still billions left to be connected, mainly in the developing world). It is, in fact, precisely this mechanism that underlies the technological dualism that exists between rich and poor countries and within the latter between rural/periurban and urban sectors. And indeed, especially the rural areas in the latter countries had seen very little major innovative activity in the decade prior to the arrival of the JioPhone. This, I should stress, had little to do with a lack of innovative capabilities in India. In fact, this country ‘performs well when it comes to innovation (World Economic Forum 2019, p. 6), being a good way ahead of most emerging economies and on the same level as several advanced economies’. The problem was rather that the vast majority of firms with well-developed innovation capabilities had little interest in the large, isolated, low-income market in India. Yet, for a variety of strategic reasons, Reliance Industries became heavily involved in promoting the JioPhone in just these markets. One such reason, for example, was to reduce the number of competitors in the field, by initiating a price war.4 Finally, it is well worth specifying the design changes in a KaiOS-based smart feature phone5 that distinguish it so sharply from a high-end smartphone, such as Apple’s iPhone. To quote Ovide (2019) on this, in relation to KaiOS partners MTN and Orange in Africa, The body of a KaiOS phone is as basic as it gets. There’s no touchscreen, which tends to be the priciest smartphone component and a battery hog. The models that Orange sells . . . have a screen that’s less than half the size of the latest iPhones and controlled with an old-school keypad. The keys are made from the least sensitive plastic possible . . . To save money, KaiOS also shrank the memory to about one quarter or less that of the cheapest Android smartphone. That means the phones can handle only one task at a time . . . For some KaiOS models, Qualcomm Inc. refashioned an old version of its processor, the phone’s brain, at an estimated cost of about $3, compared with the roughly $50 version found in top-end smartphones. In total, KaiOS-powered phones are made from about $15 worth of parts–Apple Inc.’s top of the line iPhone has $390 worth of stuff (Ovide 2019, emphasis added).

This is not at all to say that the excess parts in the Apple phone are always wasted. The point is rather that they provide services that are valued by those with high rather than low incomes. And for reasons that are explored next, it seems to be relatively poor Indians who most value the more appropriate version of mobile Internet (and specifically smart feature phones).

4

Most obviously by selling the JioPhone for effectively nothing. Note that the use of open-source (OS) technology itself reduces the price of the smart feature phone.

5

4.5 Adoption

4.4

41

Pre-Adoption (Binary Decisions)

I have added this phase to highlight the binary variables that determine whether or not the sequential process will actually proceed to the adoption stage (GSMA 2020). On the one hand, there is the question of whether mobile broadband coverage is sufficient to allow adoption on the part of those wishing to use the technology. If not, no other determinants of adoption will be able to compensate for this pre-adoptive variable and the entire process will be halted for those who are unable to gain coverage.6 Similarly, a lack of Internet awareness will render irrelevant the values of other variables in the sequence described in Fig. 4.1, because they cannot substitute for a lack of awareness of the new technology and once again the process will be aborted at this early stage. In India, however, neither of these extreme outcomes seems to have posed a serious threat. Indeed, with respect to 4G, coverage is almost complete (GSMA 2020, p. 14). The result, partly, of the ‘most significant expansion in mobile broadband coverage in 2019 . . . as Reliance Jio covered almost 99% of the population with its 4G network . . . surpassing both 2G and 3G coverage’ (GSMA 2020, p. 40). There was, that is to say, a clear case of leapfrogging as the country bypassed 2G and 3G options in favour of the 4G alternative. By way of comparison, it is useful to consider that in 2019 4G covered 82% of the population of low- and low-middleincome countries, of which India forms a part. From this point of view too, therefore, the country not only performs quite a bit better than its peers, but also (for this reason) poses no threat to adoption. As for the other component of this phase, awareness of the Internet, I was unable to find any data whatever. However, this too does not seem to me to represent a serious impediment to reaching the adoption stage and beyond. For one thing, its own remarkable success has contributed to making the JioPhone into something of a household name in India. To this should be added the formidable advertising campaigns conducted by Reliance Jio, featuring one of India’s best-known film stars (Techcrunch 2020). More precisely, Reliance Jio has used a variety of channels to place its advertising messages, namely on ‘television, radio, newspapers, magazines, and billboards, as well as social media platforms, including Instagram, Facebook, Twitter and YouTube’ (Padmaja and Antony 2013).

4.5

Adoption

If the outcome of the previous stage is one that allows widespread adoption in principle, the question is still whether the relevant variables favour that outcome or not. In Fig. 4.1, I do not purport to discuss all such variables, but rather the ones that 6

In other words, coverage is a necessary condition for adoption.

42

4 Why Is India So Dominant in the Demand for New Smart Feature Phones That. . .

appear from the literature to be most salient for those with relatively low incomes in developing countries. On this basis and in specific relation to low- and low-middle-income countries, the latest GSMA report (2020), for example, opines that ‘Affordability of handsets remained the top reported barrier to mobile ownership in many LMICs in 2019’ (GSMA 2020, p. 22). To these costs need to be added those that arise from the cost of data, which is also frequently reported as a major determinant of adoption (Wang 2020). It bears emphasis in this context that both types of costs are especially important in countries and regions where incomes are particularly low and those living in poverty may find that even the relatively low costs of a smart feature phone are beyond their means. For example, ‘According to a recent survey, in Africa on average 1GB of mobile data costs 8% of monthly income, well above the ‘1 for 2’ affordability threshold, where 1GB of mobile data should cost no more than two per cent of the average income’ (Channel Vas 2021). Mobile phone costs, moreover, include the price of a handset. For example ‘In 2018, the median cost of an entry-level internet device in Africa was 40% of monthly income, and the mean average was 62% of monthly income. For the poorest 20% of the population, the average cost of a device in Sub-Saharan Africa was 375% of monthly income in 2018’ (Channel Vas 2021). At the other end of the spectrum, by contrast, lay India, with the lowest costs in the world, in the case of both handsets and data. With regard to the former, I have already referred to the heavy subsidies granted to the JioPhone by Reliance Industries, which allowed the product to be sold for virtually nothing, while the cost of data was said to be the lowest in the world at a price of US$0.26 for 1GB (The Economic Times 2019). All this meant that the Jio was affordable even to some of the poorest Indians in rural and other marginalized areas. And a good case can be made, furthermore, that ultra-low costs were the primary reason why India has been so dominant a user of the new smart feature phone. The more so, in fact, because according to the GSMA, costs are the main barrier to ownership in many low- and low-middle-income countries. And especially to the vast numbers of users who were priced out of the market by the much higher cost of smartphones in India and elsewhere. The impact of the exceptionally low costs just referred to, on the adoption of the JioPhone, would however have been much muted by a weak performance on one or more of the other major determinants of adoption. I am referring here especially (but not only) to the country’s availability of literacy and digital skills. For, according to the GSMA (2020), ‘A recent study showed that among those who are aware of mobile Internet, a lack of literacy and digital skills was the biggest perceived barrier to adoption across regions, with 34% of people highlighting this as the reason for not coming online’. The severity of the problem, in specific cases, however, will depend on the type of skill concerned and the backwardness of the country with regard to the availability of that particular skill. Of the three types of skill identified by the ITU (2018), referred to as basic, intermediate, and advanced, the first is likely to comprise the least serious constraint, because its goal is merely to ‘enable us to function at a minimum level in

4.5 Adoption Table 4.1 A selected sample of low-middle-income countries, digital skills, 2019

43 Country Vietnam Algeria Angola Bangladesh Benin Bolivia Cambodia Cameroon Egypt India Nepal Pakistan Tanzania Senegal Nigeria Philippines Tunisia Zambia Average

Digital skills (7 is best) 3.77 3.97 2.45 3.55 3.68 3.22 3.57 3.9 4.66 4.43 3.67 4.15 3.87 4.21 3.42 5.06 4.23 3.5 3.85

Source World Economic Forum (2020), Global Competitiveness Index

society’ (ITU 2018). Nevertheless, even the acquisition of basic and intermediate skills in low- and low-middle-income countries demands serious attention on the part of governments, which, all too often, are loathe to take the issue seriously enough, preferring instead to focus on the mere provision of access to digital technology (World Bank 2016). My specific concern, however, is with the Indian performance on digital skills as compared to a selected sample of lower-middle-income countries. Fortunately, recent data are available as shown in Table 4.1. The results show that India is one of the leading countries in the group with a score that is well above the average of 3.85. Taken together, therefore, the costs of the JioPhone and the availability of digital skills appear consistent with the rapid and extensive diffusion of the product in India (as too does the almost complete coverage of 4G in the country). There is still much, however, that determines the welfare impact of the smart feature phone, and it is to these post-adoption factors that I turn next, following Fig. 4.1.

44

4.6

4 Why Is India So Dominant in the Demand for New Smart Feature Phones That. . .

Use of Adopted Smart Feature Phones

It is well to note from Fig. 4.1 that literacy and digital skills appear not just as determinants of adoption, but also as variables that influence the use of the JioPhone in India once it is adopted. My focus here includes isolated rural and peri-urban areas because it is there that the shortage of digital skills tends to be most acute (ITU 2018). It is also these areas therefore that tend to derive, at the use stage, the least benefits of smart feature phones, though they are most in need of them.7 Several recent publications have noted for example that in India and other developing countries, the pattern of Internet use tends to be limited and biased in favour of entertainment, rather than, as one might expect, developmental activities (Wang 2020).8 It is, of course, possible that the observed patterns of use do indeed reflect the ‘true’ preferences of low-income inhabitants. But it is just as likely that the revealed preferences are formed on the basis of inadequate information and an acute lack of digital skills. Wang’s (2020) case study of Nigeria, for example, found that even those who claimed to be regular Internet users stuck nevertheless to the most rudimentary online activities. For new users, who comprise the bulk of smart feature phone owners, lack of knowledge also limits the welfare gains from these technologies (as also emphasized in the next section, where the same variable operates). Wang concludes that ‘Without exposure to activities like downloading apps, setting up online accounts, using web browsers, or making video calls, new users never learn all the internet has to offer them’ (Wang 2020). What also bears on the use of smart feature phones in India is the extent of local language support and country-specific services (which again influence the derived welfare effects). Regarding the former, ‘an advantageous factor’ with the JioPhone in particular is the unusually high number of, twenty-two, Indian languages that it supports. Meanwhile, the Voice Assistant on that phone can respond to commands in both Hindi and English, which should not be underestimated in a place where a majority of the population is uncomfortable with English (Firstpost 2021). As regards country specificity, moreover, the appeal of the Jio was undoubtedly boosted by access to major social media applications such as YouTube, WhatsApp, and Facebook. ‘And its not just international apps being made for the KaiOS platform that will differentiate one KaiOS phone from the other. The trick lies in developing apps which are relevant to the Indian milieu. JioPhone has an app like Jio Kisan which lets farmers sell their products via the mobile phone. Voice commands in Hindi can be super valuable for the elderly or those not comfortable with typing out commands (such as, most notably, the vast numbers of illiterates) (Firstpost 2021, emphasis added). WhatsApp, too, offers this advantage to the

7

This is obvious just from literacy data by country. See World Bank data. Such activities include job search, use of health and farming sites, and those that offer online courses on digital literacy. In fact, the Indian Prime Minister initiated a digital literacy programme that was intended to reach 60 million households. But, so far as I am aware, it has not yet been formally appraised.

8

4.8 Conclusion

45

groups mentioned, because the voice communication that it offers substitutes for the text that is more or less inaccessible to them.

4.7

Welfare Effects

As noted above, the welfare effects of the smart feature phone depend on the outcomes of earlier stages in the process. Some of these are unambiguously positive such as the use of voice rather than text, by illiterate and elderly populations. So too is access to local and relevant applications and the ability to communicate more easily with friends and family by means, say, of e-mail. Assessing other welfare effects, however, is a much trickier proposition. As shown in the last column of Fig. 4.1, for example, there is an item labelled ‘time spent on the Internet’. And the evidence shows that in general relatively poor countries spend more time on this than rich ones. At first glance, this might suggest that the former derive more benefits from the Internet than the latter. On closer inspection of these results, however, one may reach a quite different welfare conclusion. For one thing, the extra time spent may be based on severely inadequate information which leads to the wrong conclusions. Or, because of a lack of literacy and digital skills, preferences may favour entertainment rather than development. This is sometimes referred to as a ‘shallow’ use of the Internet. Then again, a very popular use of this technology in developing countries is gaming, which in some, if not many, circumstances can be regarded as a welfare gain. In other cases, however, serious addictions develop, often with disastrous consequences. Then, too, there is the problem that the product that the buyer most prefers is not available. Here as well, therefore, it is inaccurate to state that the choice revealed by such a person is necessarily his or her optimal one. What I conclude is that in developing country circumstances, which usually include highly imperfect information and a lack of digital skills, more needs to be done to examine the environment in which the Internet is being used. Up till now the focus has been largely to promote the mere adoption of the technology without examining the circumstances in which it is selected and used. Yet, these circumstances play an important role in determining the welfare effects of the new low-cost smart feature phone.

4.8

Conclusion

Using a multi-sequential analytical framework, this chapter set out to investigate the reasons why India has dominated the market for the new low-cost Internet-enabled smart feature phones in recent years. Indeed, of the total demand for these phones in

46

4 Why Is India So Dominant in the Demand for New Smart Feature Phones That. . .

2021, the highest share is expected to come from India (Counterpoint Research 2019). The proximate cause of this outcome seems to have been the ultra-low costs associated with the JioPhone in India, not only in terms of the effectively zero price of the handset, but also the price of data, which, at least in the early years of the project, was the lowest in the world. At these extraordinary prices, the Jio was accessible to even some of the poorest rural groups, a rare occurrence even with innovations that are expressly designed for such groups. Mainly for that reason, as I see it, the Jio had reached sales of 100 million units in India by 2020 and much lesser amounts in other countries, mainly in Africa, in conjunction with the biggest telecom operators there, MTN and Orange (though these phones, too, use the KaiOS OS and were sold for around US$20). The more fundamental cause, however, can be traced back to historical circumstances that were distinctly coincidental in nature. For, on the one hand, although the OS firm Mozilla had abandoned hope of using Firefox as a basis for smartphones, it appeared a few years later in modified form as the OS for the JioPhone in India. And KaiOS, in turn, because of its earlier experience with an Alcatel brand, was able to win a contract with Reliance Industries, India’s biggest firm. This turned out to be a highly profitable quasi-coincidence, because of the subsequent success of the JioPhone and the desire on the part of both firms to fill the glaring gap between basic mobile phones (without Internet connectivity) and expensive smartphones. This is not to say, though, that KaiOS-based phones can never be successful elsewhere on a lesser scale, since there were factors associated with the Jio that also contributed to its success other than vast subsidization. Design changes, for example, have allowed KaiOS-based smart feature phones to be sold in sub-Saharan Africa at an average price of around US$20, with what, at an early stage, seem like promising results. But without the support of a huge collaborator such as Reliance Industries, achieving the gains will be decidedly more difficult to come by.9 Features other than prices and costs also played a role in the Jio case. Not only was there almost complete 4G coverage of India, but also with respect to digital skills, that country performed relatively well in a comparison with a sample of lowand low-middle-income nations. Then too there were numerous features which bore on local relevance and content, two important determinants of adoption and use. The Jio itself supports twenty-two local languages, while local application developers tended to design in ways that fit in with the domestic environment (although this was a tendency rather than a foregone conclusion). Jio Kisan, for example, a programme that is designed to help Indian farmers, is perhaps the best example of relevant local design at work. Note, too, finally, that the WhatsApp feature on the JioPhone is able to bring the Internet to those who are illiterate and those who are elderly. This occurred because WhatsApp was able to bring the Internet to these groups by

9

A point made to me in personal correspondence by Peter Richardson of Counterpoint (21 May 2019).

References

47

offering voice instead of text communications, which are often difficult or impossible to read. In Sect. 4.7, on welfare effects, I emphasized that while some aspects of smart feature phones are unambiguously favourable, others are much more difficult to ascertain, because they depend on the environment in which decisions are made and the technology that is used. On occasion, for example, what appear on the surface to be true preferences may instead be the product of an environment which fosters a lack of knowledge and a lack of digital skills, both partly the result of government inactivity and incompetence. In developing countries, governments often appear more reticent towards imposing product and standards than those that are more developed. Given the breadth of imperfect information that exists in the former, there is an even more compelling case for active state intervention. As I emphasize in the next chapter, however, state failure may be offset in part by initiatives emanating from the private sector. One of them, for example, is an app developed by KaiOS Technologies to provide a wide variety of useful information for first-time Internet users.

References Channel Vas (2021) Blog. The challenge of affordability of devices in Africa. Available https:// channelvas.com/article/the-challenge-of-affordability-of-devices-in-africa Counterpoint Research (2019) Smart feature phones to create US$28 billion revenue opportunity. Available https://www.counterpointresearch.com/smart-feature-phones-create-us28-billion-rev enue-opportunity/ Firstpost (2021) JioPhone and KaiOS are redefining the feature phone landscape in India. Available https://www.firstpost.com/tech/news-analysis/jiophone-and-kaios-are-redefining-the-featurephone-landscape-in-india-4681121.html GSMA (2020) The state of mobile connectivity. Available https://www.gsma.com/r/wp-content/ uploads/2020/09/GSMA-State-of-Mobile-Internet-Connectivity-Report-2020.pdf ITU (2018) Digital skills toolkit. Available http://handle.itu.int/11.1002/pub/8110cd77-en James J (2018) A sequential analysis of the welfare effects of mobile phones in Africa. Soc Sci Comput Rev 37(2):279–290 Livemint (2019) Reliance JIO sold 5 crore smart feature phones in less than 2 years. Available https://www.livemint.com/technology/tech-news/reliance-jio-sold-5-crore-smart-featurephones-in-less-than-2-years-report-1550635450416.html Ovide S (2019) The next big phones could bring a billion people online. Bloomberg Businessweek. Available https://www.bloomberg.com/news/features/2019-06-07/the-next-big-phones-couldbring-a-billion-people-online Padmaja K, Antony P (2013) Leverage of e-marketing: a case study of Reliance JIO. Available https://www.coursehero.com/file/74767076/20-Leverage-of-E-Marketing-A-CaseStudy-of-Reli ance-Jiopdf/ Schumacher EF (1973) Small is beautiful. Blond & Briggs Techcrunch (2020) KKR to invest US$1.5 billion in Reliance Industries’ Jio platforms, its biggest deal in Asia. Available https://techcrunch.com/2020/05/21/kkr-invests-1-5-billion-in-indiasreliance-jio-platforms/

48

4 Why Is India So Dominant in the Demand for New Smart Feature Phones That. . .

The Economic Times (2019) Mobile data price in India: India has the cheapest mobile data in world: study. Available https://economictimes.indiatimes.com/tech/internet/india-has-the-cheapestmobile-data-in-world-study/articleshow/68285820.cms Wang Y (2020) First time internet users in Nigeria use the internet in a unique way. Here’s why that matters. Available https://www.kaiostech.com/first-time-internet-users-in-nigeria-use-the-inter net-in-a-unique-way-heres-why-that-matters/ World Bank (2016) World development report: digital dividends. Available https://www. worldbank.org/en/publication/wdr2016, Washington, D.C World Economic Forum (2019) The global competitiveness report. Available http://www3. weforum.org/docs/WEF_TheGlobalCompetitivenessReport2019.pdf World Economic Forum (2020) Global competitiveness report. Available https://www.weforum. org/reports/the-global-competitiveness-report-2020 Zanello G, Fu X, Mohnen P, Ventresca M (2015) The creation and diffusion of innovations in developing countries: a systematic literature review. J Econ Surv 30(5):884–912

Chapter 5

Interregional and Intercountry Analysis of Mobile Internet Connectivity in Sub-Saharan Africa

5.1

Introduction

Sub-Saharan Africa is well known as being the poorest region on the globe and as such is least well placed to reap the many advantages of mobile Internet connectivity. This is because low income exerts a negative influence on many of the major determinants of connectivity such as coverage, affordability, and education (e.g. literacy and digital skills). Yet, recent years have seen a marked improvement in mobile Internet connectivity across much of the region, as a result of positive changes in some of the main determinants of this variable (such as affordability and digital skills). Silver and Johnson (2018), for example, have noted that ‘Sub-Saharan Africa has experienced dramatic gains in internet use in recent years. With this rapid growth in connectivity . . .’ (2018, n.p.). In order to better understand this performance, my first task is to establish which determinants of Internet connectivity have improved in recent years and to determine, by means of an interregional analysis, if sub-Saharan Africa has in some dimensions now reached the status of some other world regions. Further understanding of Internet connectivity in Africa is sought through comparisons at the country rather than the regional level. That is to say that according to most recent data, countries are ranked according to the main determinants of connectivity described in Chap. 4. Per capita income, for example, goes some way towards an explanation, because of its influence on coverage, affordability, education, and local content, but other factors (such as government policy) also need to be taken into account. It is to this interregional analysis that I first turn, after showing a ranking of regions according to mobile Internet connectivity, which, in a sense, is the dependent variable of the analysis.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. James, New Perspectives on Current Development Policy, SpringerBriefs in Economics, https://doi.org/10.1007/978-3-030-88497-0_5

49

50

5.2

5 Interregional and Intercountry Analysis of Mobile Internet Connectivity in. . .

Interregional Analysis of Determinants of Mobile Internet Connectivity

Table 5.1 provides data (in descending order) on regional connectivity rates. Not surprisingly, sub-Saharan Africa, as the poorest region, is the least connected and at the other extreme lie Europe and East Asia, with rates of 70% and 60%, respectively. Note also the increase in the percentage of the population in sub-Saharan Africa that is connected to the mobile Internet. Thus, beginning with 13% in 2014, the share grew in the years between 2015 and 2019 at the consecutive rates of 15%, 18%, 20%, 23%, and 26% (GSMA 2020b). I begin with income differentials in Table 5.2, because even a cursory comparison of that table with Table 5.1 indicates a close correlation between the two. In fact, there turns out to be a perfect rank correlation between income and connectivity (using the Arab world instead of MENA). That is to say, that the ordering of regions is the same in both tables, which is due partly to the effect of income on many of the other variables that determine connectivity, such as broadband coverage, affordability of handsets and data, education, and so on. Together with the way in which it is distributed, for example, income determines the extent of poverty in a region, which is shown in Table 5.3 (with some change in the definition and inclusion of regions). Note too that the correlation between income and connectivity does not in itself show that the causation runs entirely from the former to the latter. For it could just as well run in the other direction, as greater connectivity enables a host of Internet

Table 5.1 Regional rates of Internet connectivity Region Europe East Asia Latin America Middle East and North Africa (MENA) South Asia Sub-Saharan Africa

Connectivity rate (%) 70 60 54 43 33 26

Source GSMA (2020a) Table 5.2 Per capita incomes across regions, 2019 (current prices in US$) Region East Asia and Pacific Latin America (excl. High income) Arab world South Asia Sub-Saharan Africa Source Statista (2020)

Per capita income 2019 11,526.74 8847.43 6580.06 1959.92 1585.44

5.3 Affordability: Handsets, Data, and Incomes

51

Table 5.3 Regional poverty estimates (using 2011 PPP and US$1.9 a day poverty line) Region Europe and Central Asia East Asia and Pacific Latin America and Caribbean MENA Sub-Saharan Africa

Headcount (%) 1.13 1.18 3.8 7.22 40.19

Number of poor (millions) 5.57 24.55 24.2 27.98 433.39

Source World Bank (2018), by license CC-BY-4.0-IGO

activities that raise productivity and enhance growth of incomes. Only an econometric analysis can unravel these two directions of causality. What is striking about the contents of Table 5.3 is not that there is again a close rank correlation of regions in terms of income, poverty, and connectivity, but rather how much more poverty there is in sub-Saharan Africa than in the other regions. For, as I shall soon demonstrate, such widespread poverty makes it extremely difficult for handset and data costs to be affordable for more than a relatively small proportion of the population. Nonetheless, it needs to be recognized that some African countries have been among the fastest growers in the world in recent years and that this must to some extent have caused a reduction in poverty and an increase in Internet connectivity. I am referring here especially to East Africa, where growth has been notably pronounced. Moreover, according to researchers at the Brookings Institution, ‘many countries are making progress towards ending poverty, including in Sub-Saharan Africa. Today, four countries already have poverty rates of below 3%: Equatorial Guinea, Gabon, Mauritius, and Seychelles. Currently, Mauritania and Gambia are projected to join this group by 2030’ (Hamel et al. 2019, n.p.).

5.3

Affordability: Handsets, Data, and Incomes

In the discussion that follows, affordability is defined as the price of an entry-level Internet-enabled phone divided by monthly GDP per capita (GSMA 2020a, p. 24). Table 5.4 lists the regions according to this definition. In what is by now a familiar pattern, sub-Saharan Africa occupies the last place in the table, due, in this case, either to a relatively high cost of handset or a low per capita income, or both of these. What needs to be emphasized, though, is that the gap in affordability with other regions has fallen in recent years on account of income growth (noted above) and a reduction in the price of Internet-connected handsets (due, in part, to the introduction of smart feature phones). As regards the latter, for example, the average affordability of the cheapest Internet device (as a percentage of monthly GDP per capita) has gone from 57% in 2015 to 46% in 2017 and finally to 30% in 2019. Such a rapid reduction was

52

5 Interregional and Intercountry Analysis of Mobile Internet Connectivity in. . .

Table 5.4 Affordability of an entry-level Internet-enabled phone by region, 2019 (ordered by most affordable region)

Region Latin America and Caribbean Europe and Central Asia MENA East Asia and Pacific South Asia Sub-Saharan Africa

Affordability 9.4 9.9 15.3 18.5 20.9 30.2

Source GSMA (2020) Note Affordability is a combination of income and device price. In the table the figures refer to the average affordability of the cheapest Internet device as a percentage of monthly per capita income

certainly due in part to the introduction in many African countries of the so-called smart feature phone, which is roughly a hybrid of traditional feature phones (without Internet connectivity) and the much more sophisticated and expensive smartphones, made largely in and for the developed world. The former product, on the other hand, ‘rethought everything to keep the essential capabilities of the smartphones but strip out costs and preserve battery life for people who likely have spotty access to electricity’ (Ovide 2019). Note, too, that the smart feature phone is based on opensource software, originally developed by Mozilla.1 For these reasons, the new technology sells in numerous African countries at or around US$20, a price that is a good deal cheaper than the typical smartphone. But though it must therefore have contributed to greater affordability in the region, for the poorest 20% of the population the new technology still represents an unduly high percentage of monthly per capita income. Thus, ‘In addition to reducing the cost of handsets, other solutions to increase affordable access of devices need to be considered. During the past few years, many consumers in LMICs [low and middle-income countries] that could not afford to purchase a phone in a single upfront payment have benefitted from asset financing models (such as payment instalment plans, subsidies, loans, leases or rentals)’ (GSMA 2020a). The affordability of mobile Internet use, however, depends not only on handset costs, but also on the costs of data. And in this respect, too, sub-Saharan Africa is the least affordable region among those shown in the previous table (defined as the highest monthly cost of 1GB as a proportion of income). But again, in common with handset costs, recent years have on the whole witnessed a sharp decline in data costs in sub-Saharan Africa. In particular, while 1GB of these costs amounted to 8% (as a percentage of monthly GDP per capita) in 2015, two years later the percentage had fallen to 5.4 and by 2019 it had reached 4.2 (GSMA 2020a). In fact, 2019 ‘marked the first year that there were more mobile broadband connections than 2G, as consumers benefitted from more affordable smartphones and smart-feature phones’ (GSMA 2020a).

1

For a detailed description, see James (2020).

5.4 Digital Skills Table 5.5 Digital skills by region 2020 (7 is highest)

53 Region USA Europe Asia Africa Latin America

Digital skills 5.5 4.7 4.7 3.7 3.6

Source World Economic Forum (2020a)

5.4

Digital Skills

Digital skills are often described as the main barrier to connectivity in sub-Saharan Africa, though they are arguably the most difficult to analyse because of the variety of definitions of them that are available. For the regional comparison shown in Table 5.5, I adopt the data provided by the World Economic Forum, in its latest Global Competitiveness Index.2 Because of the striking result that the measure of digital skills in sub-Saharan Africa is the same as in Latin America, a richer and more developed region, I also sought comparative data from the GSMA. Unfortunately, this institution only looks empirically at the issue from the perspective of countries, rather than regions. Nevertheless, by aggregating data from the former level to the latter, I was able to compare digital skills measures of three regions, namely, sub-Saharan Africa, Latin America, and Asia. Again on a scale of 1 to 7, the scores are, respectively, 3.65, 3.56, and 4.32. In this case, therefore, sub-Saharan Africa does not just do equally as well as Latin America, but actually exceeds it (albeit by a small margin). Part of the explanation would seem to be that the former region contains a few more extreme outliers than the latter. I am referring here particularly to Kenya (4.55), Ghana and Senegal (4.21), and Rwanda (3.96). That these countries stand out is to be expected, since they are known to pay particular attention to (information) technological issues. For example, ‘Ghana has been “nimble” in addressing the opportunity that digital presents, according to market participants, and it has implemented these technologies in education policies’ (IFC 2019). Other studies, however, point to different countries when ranking digital skills across sub-Saharan Africa. For example, Madden and Kanos find that ‘citizens of Nigeria, Kenya and South Africa have a higher level of digital skills than the rest of sub-Saharan Africa on average’ (Madden and Kanos 2020). And speaking of South Africa, the same authors note that the ‘country’s relative penetration of digital skills is slightly above the global average’ (Madden and Kanos 2020). Furthermore, insofar as technological hubs represent the acquisition of digital skills, Table 5.6 shows that this country is far ahead of others in the region (in fact, such hubs can be viewed as a determinant and outcome of digital skills).

2

And at the same time to accept the limitations of that index, not the least of which is that it is subjective (see World Economic Forum, Global Competitiveness Index (2020a)).

5 Interregional and Intercountry Analysis of Mobile Internet Connectivity in. . .

54

Table 5.6 Technological hubs in sub-Saharan Africa, 2016

Country South Africa Kenya Nigeria Ghana Senegal

Number of hubs (from most to least) 54 27 23 16 10

Source GSMA, Data System Accelerator, cited in ECDPM (2016) Table 5.7 Rural–urban gap in mobile Internet use, 2018

Region East Asia and Pacific Latin America and the Caribbean Europe and Central Asia South Asia MENA Sub-Saharan Africa

% 23 29 26 45 36 58

Source GSMA (2019) Note The term rural–urban gap refers here to how much less likely a person living in a rural area is to use mobile Internet than a person living in an urban area

Note, finally, that the acquisition of digital skills influences other variables, not the least of which is digital adoption. To cite the World Bank, Knowledge/skills impact digital adoption in two primary ways. First, the knowledge/skills of end users, such as farmers, affect demand for digital technologies. For example, farmers have to know that a certain technology exists, believe it will help them . . . and learn how to use it. Second, the knowledge/skills of entrepreneurs that develop digital technologies impacts digital adoption (the correlation coefficient of basic skills and mobile applications developed per person was 0.89 using 2017 cross-country data) (World Bank 2019).

5.5

Rural–Urban Gaps by Region

Cutting across the variables that have been discussed above is the rural–urban gap, which, as Table 5.7 shows for 2019, is by far the most severe in sub-Saharan Africa, at 58%. The problems are really twofold. One is that conditions in rural areas are generally unfavourable from the point of view of the factors discussed above, such as digital literacy and general literacy, affordability, coverage, and relevance. Much of this is due, in turn, to lower incomes and higher poverty in rural rather than urban areas. Geographical factors too play a role in the form of isolated villages which have only a remote connection to digital technology and the Western-based content that is

5.6 Intercountry Variations in Internet Connectivity and its Determinants Table 5.8 Mobile broadband connections (% penetration 2019, and per capita incomes 2019)

Country South Africa Mauritius Botswana Ghana Côte d’Ivoire

Penetration (%) 125 104 97 90 87

55 Per capita income 6001.9 11,099.2 7961.3 2202.1 2276.3

Source GSMA (2019, p. 51) and World Bank data (2019), by license CC BY-4.0 Note Penetration rates can exceed 100% because they refer to SIM cards rather than devices

mostly provided on the Internet. The second problem is that the rural areas are much more predominant in sub-Saharan Africa than other developing regions.3 Before I turn to intercountry comparisons in the regions, it is well to consider the main results that have thus far been reported. The main one, perhaps, is that despite rapid recent increases in Internet connectivity, sub-Saharan Africa still lags behind other regions. Indeed, there is only one case where the region has achieved a score that is equal to one of the others. But because there are often marked differences in country performance within regions, it is instructive also to consider outliers at this lower level of aggregation. According to data contained in a recent Brookings report, for example, ‘citizens of Nigeria, Kenya and South Africa, have a higher level of digital skills than the rest of Sub-Saharan Africa on average’ (Madden and Kanos 2020). Or, to take another example, Mauritius and Seychelles have a youth male literacy rate of 99, as against the value of 79 for the region as a whole (World Bank, 2020).

5.6

Intercountry Variations in Internet Connectivity and its Determinants

In the tables that follow, I list the five best-performing countries on the variables I selected in the introduction to the chapter, beginning with mobile broadband connections (% penetration). I also include in the table per capita incomes (Table 5.8). As it did in much of the interregional analysis, income per capita tends to correlate with other variables. In this case, the top three countries have substantially higher incomes per capita than the next two, although these demonstrate that relatively high percentage penetration can be achieved even by countries with modest incomes. This, indeed, is a recurring theme throughout the chapter, and it is linked in many cases to an active government policy towards digital technology. The role of income is also prevalent with regard to the local relevance of mobile Internet. As shown in Table 5.9, for example, the top three places are occupied by

3

On this see World Bank data on percentage of rural population by region.

56

5 Interregional and Intercountry Analysis of Mobile Internet Connectivity in. . .

Table 5.9 Local relevance of mobile Internet

Country Mauritius South Africa Gabon Botswana Kenya

Penetration (%) 71.8 61.1 57.8 47.1 45.4

Source GSMA (2020c) Table 5.10 1 GB data costs (cheapest first)

Country Somalia Sudan Tanzania Ghana Kenya

Cost (US$) 0.50 0.63 0.73 0.94 1.05

Source Statista (2021)

Mauritius, South Africa, and Gabon, all countries with relatively high incomes relative to the rest of sub-Saharan Africa. The appeal of income in this case seems to be due mainly to the fact that local developers and telecom operators find it more profitable to produce for markets with higher rather than lower incomes. And in order to supply these markets, such firms need to provide local in addition to global content.4

5.6.1

Affordability

Affordability, as noted above, comprises handset and data costs relative to income. And the second component, data costs, provides the most severe challenge to the notion that income invariably underlies the ranking of regions as well as countries. Consider, in this regard, the entries in Table 5.10. Thus, although sub-Saharan Africa on the whole has by far the most expensive data costs of all regions (as noted above), the cheapest countries in the region are among the poorest in the world and the others are described by the World Bank as low-middle income. In relation to the most anomalous case, Somalia, Gilbert (2020) has observed that ‘Telecom services have grown substantially . . . over the past ten years and low data prices may be because of the strong competition in the market with at least 11 different providers operating locally’ (Gilbert 2020, n.p.).5 4

On the other hand, it could also be argued that low-income inhabitants tend to be more interested in local, rural issues than those that are international in character. 5 The Tanzania case, too, has been ascribed to intense competition among providers (CGTN Africa 2016).

5.6 Intercountry Variations in Internet Connectivity and its Determinants Table 5.11 Mobile phone prices and affordability (from lowest to highest)

Country Lesotho Mozambique Guinea Mali Botswana

Price (US$) 17.96 18.76 20.70 25.29 26.02

Country Botswana Namibia Mauritius Gabon Lesotho

57 Affordability (%) 4.03 7.81 8.27 8.93 15.62

Source Alliance for Affordable Internet (2020)

The presence of Tanzania and Kenya in the table should not be seen in terms of these countries alone, but rather from the point of view of East Africa more generally. For, as numerous studies have found (e.g. RIA), this region has the lowest Internet data cost in the entire African continent. Other countries in the region with low data costs include Rwanda, Uganda, and Burundi. Whether this is a mere coincidence or due to the presence of a common factor in the group is not clear. But it is certainly a question that deserves an answer, because of the different policy lessons that flow from it. Some see the answer in common policy reforms that have been undertaken in East Africa, such as the strengthening of institutions, improving policy coordination, improved ease of doing business, and so on. ‘Some countries–particularly the smaller economies of East Africa—are already demonstrating how powerful such reforms can be. If the entire continent took this approach . . . some believe that Africa could emulate China’s rapid rise of the last 50 years’ (World Economic Forum 2020b, n.p.).6 Economic growth in East Africa has certainly been very rapid, especially in countries such as Ethiopia, Rwanda, Tanzania, and Kenya. As with data costs, the price and affordability of handsets are a major determinant of adoption of Internet-enabled phones in sub-Saharan Africa. And again, the data shown in Table 5.11 do not (at least with respect to price) suggest that income has much to do with it. It is true that Botswana, in last place, falls into the category of high-middleincome countries, but Lesotho is in the low-middle group and the others are all defined as low income. It is only when income is introduced in order to calculate affordability that relatively high-income countries such as Mauritius, Namibia, and Gabon enter the picture.

5.6.2

Network Coverage

It is noteworthy that as in two previous Tables 5.8 and 5.9, Mauritius and South Africa occupy the first two places. That it occurs on three different occasions points to the striking role of income as a determinant of Internet connectivity in the 6

It is certainly true that until 2019 East Africa was growing faster than any other region on the continent, led by Ethiopia, followed by Rwanda and Tanzania (African Development Bank 2019).

58

5 Interregional and Intercountry Analysis of Mobile Internet Connectivity in. . .

Table 5.12 Network coverage, 2019

Country Mauritius South Africa Rwanda Senegal Botswana

Score (%) 89.1 88.1 85.7 77.3 75.8

Source GSMA (2020a) Note Score has been normalized to have a value between 0 and 100, with a higher score representing stronger performance

region. What is surprising, though, is that Rwanda, as a lower income country, appears in third place in the table, showing once again that income is by no means the only explanation of Internet connectivity differences in sub-Saharan Africa. In the case of Rwanda, for example, an ITU report notes the role of government policy in the country’s experience. Thus, ‘Government steps of creating a regulator, privatizing the incumbent, introducing competition and developing a broadband strategy have resulted in a high level of broadband infrastructure in the country. The SMART Rwanda Master Plan envisions taking ICTs to the next level by using their transformational capabilities and developing a vibrant ICT-enabled sector’ (ITU 2018, p. 149/150) (Table 5.12). This case is of great relevance to the many other low-income countries in the region that are seeking to raise their rates of Internet connectivity, without involving themselves in excessive spending. But what the Rwandan case also indicates is that a strong political will is needed in order to effect a major digital policy change. And this, unfortunately, is a quality in scarce supply in much of sub-Sahara.

5.6.3

Digital Skills

The lack of digital skills is often cited as the main barrier to Internet adoption in developing countries, especially in those with relatively low incomes. In fact, according to Kandri (2019) ‘Nowhere is the need for digital skills more pressing than in Africa . . . To ensure positive economic growth and development that benefits all citizens, Africa must make big strides in digital literacy. In Sub-Saharan Africa, 230 million jobs will require digital skills by 2030, according to the Digital Skills in Sub-Saharan Africa report that IFC recently published . . . The status quo is unacceptable. Only 50 per cent of countries in Africa have ‘computer’ skills as part of their school curriculum, compared to 85 per cent of countries globally’ (Kandri 2019, n.p.). This raises the salient question of which countries in the region have performed exceptionally well with regard to digital skills more generally. Part of the answer to this question is contained in the results shown in Table 5.13.

5.6 Intercountry Variations in Internet Connectivity and its Determinants Table 5.13 Leading performers on digital skills, by country (7 is best), 2019

Country Kenya Mauritius Senegal Ghana Gambia

59

Digital skills score (7 is best) 4.5 4.3 4.2 4.2 4.0

Source World Economic Forum (2019), Global Competitiveness Index dataset Table 5.14 Youth literacy (%)

Country Mauritius Botswana South Africa Namibia Eswatini

Youth literacy (%) 99 97 95 95 95

Most recent year 2018 2013 2017 2018 2018

Source World Bank (2020), by license CC-BY-4.0 Note Botswana refers to an earlier year

Mauritius is yet again in the top two places, indicating how substantial and broad is the country’s contribution to the factors that influence Internet connectivity. Kenya’s exceptional performance is also not surprising since the country has a long tradition of promoting technological development in general and information technology in particular.7 Something of the same tradition is also to be found in Ghana (which, one may recall, had the second highest number of technology hubs in sub-Saharan Africa). The third place in that ranking was occupied by Kenya, which too excels in this dimension. In the context of digital skills it would be remiss of me to say nothing about literacy, which is a requirement of some such skills. The first thing to say is perhaps that there are sharp differences between country clusters, according to the World Bank definition of high, low, and middle incomes. According to UNESCO, for example, ‘In low-income countries, the rates of children and adolescents not learning are systematically higher than in lower-middle-income, upper-middle-income and high-income countries’ (UNESCO 2017, p. 15). In terms of reading, for example, there is evidence that shows the percentage of the school-age population that will not attain minimum proficiency levels. Thus, on this measure, whereas the percentage for high-income countries was 13, for the low-income group, it was as high as 90 (UNESCO 2017). Let us next examine the ranking of countries on the literacy rates among the youth (Table 5.14).

7 ‘Kenya is regarded as the second-best innovation hub in Sub-Saharan Africa. Tech start-ups thrive in Kenya, due in part to the ready availability of credit lines and other forms of financing. 2019 was the ninth consecutive year Kenya exceeded the innovation relative to GDP figures, expected from middle-income nations’ (The Borgen Project 2020).

60

5 Interregional and Intercountry Analysis of Mobile Internet Connectivity in. . .

What this table suggests is that income plays an important role in the determination of literacy differences between countries in sub-Saharan Africa. For not only is Mauritius the only high-income country in the region, but the next three countries are all drawn from the high-middle-income category. Only Eswatini belongs to the lowmiddle-income category. There are many reasons why income should exert an influence that is inegalitarian. Not the least of them is that poor farmers in Africa’s sizeable rural areas often cannot afford to send their children to school, who are confined instead to work on the family farm. And if children and adolescents do manage to make it to school, they may not be able to afford books and other items. Then again, the schools themselves may be of low quality, with unqualified teachers who are frequently absent. (For an early econometric analysis of these variables, see Michaelowa 2001). And not least is the possibility that poor children may be undernourished, thus further impeding their chance of learning effectively.

5.7

Conclusions

This chapter has concerned itself with Internet connectivity in sub-Saharan Africa. The first part is driven by the observation that this region has performed particularly well in recent years with respect to the variables that determine Internet connectivity. By means of a series of interregional comparisons, I sought to establish whether sub-Saharan Africa has now joined the ranks of other world regions. It turns out with respect to only one variable, digital skills, does the region have the same score as another (Latin America). In all the other cases, sub-Sahara lags behind (in some instances, considerably so) other world regions. The regional analysis, however, fails to take into account that there will be other countries which perform much better than average and from which more specific lessons can be learnt. That is why the second part of the chapter was devoted to comparisons of countries within sub-Saharan Africa and to search for common patterns in the results. Overall, the best performer was Mauritius, the only high-income country in sub-Sahara, which occupies one of the top two places in many of the tables. This is not unexpected given the important role that income plays in determining coverage, affordability, education, and other variables. After all, this country enjoys much the same income level as many developed countries, for whom expensive smartphones were designed. Much the same can be said of upper-middle-income South Africa, which tends to appear in many of the same tables as Mauritius and one expects, for its relatively high income. However, there are countries which perform well on a number of indicators, where income does not play an influential role. I am thinking, here, for example, of Kenya and Ghana, where technological issues have long received serious treatment by governments, as reflected partly by the relatively high number of technological hubs that they possess. Rwanda is another

References

61

country where a strong political will to absorb new digital technologies in its development strategy is to be found. It is important though not just to consider individual countries alone, but also in terms of the regions to which they may belong. East Africa is probably the most obvious case in this regard, and according to one source, the keys to its success are ‘Its attempts at building a vibrant services sector, its coordinated state-supported industrial growth efforts, and a renewed focus on education reform with an emphasis on fostering digital skills all hold valuable lessons for the rest of the continent’ (Business Week 2020). Partly for these reasons, growth has been robust in recent years (Further Africa 2020).

References African Development Bank (2019) Regional economic outlook 2019-East Africa. Available https:// www.afdb.org/en/documents/document/regional-economic-outlook-2019-east-africa-108658 Alliance for Affordable Internet (2020) From luxury to lifeline: reducing the cost of mobile devices to reach universal internet access. Available https://a4ai.org/research/from-luxury-to-lifelinereducing-the-cost-of-mobile-devices-to-reach-universal-internet-access/ Business Week (2020) What East Africa can teach the rest of the continent about economic growth. Available https://www.busiweek.com/what-east-africa-can-teach-the-rest-of-the-continentabout-economic-growth/ CGTN Africa (2016) Tanzania ranked high in bridging economic inequality. Available https:// africa.cgtn.com/2016/10/04/tanzania-ranked-high-in-bridging-economic-inequality/ ECDPM (2016) Africa’s growing tech hubs and smart cities. Available https://ecdpm.org/greatinsights/2030-smart-engagement-business/africas-growing-tech-hubs-smart-cities/ Further Africa (2020) East Africa maintains strong economic growth. Available https:// furtherafrica.com/2020/07/20/east-africa-maintains-strong-economic-growth/ Gilbert P (2020) Somalia has Africa’s cheapest data prices, Connecting Africa. Available http:// www.connectingafrica.com/author.asp?section_id¼761&doc_id¼762763 GSMA (2019) The state of mobile internet connectivity. Available https://www.gsma.com/ mobilefordevelopment/resources/the-state-of-mobile-internet-connectivity-report-2019/ GSMA (2020a) The state of mobile internet connectivity. Available https://www.gsma.com/r/wpcontent/uploads/2020/09/GSMA-State-of-Mobile-Internet-Connectivity-Report-2020.pdf GSMA (2020b) Mobile internet connectivity. Sub-Saharan Africa factsheet. Available https://www. gsma.com/r/wp-content/uploads/2020/09/Mobile-Internet-Connectivity-SSA-Fact-Sheet.pdf GSMA (2020c) Mobile connectivity index. Available mobileconnectivityindex.com Hamel K, Tong B, Hofer M (2019) Poverty in Africa is now falling–but not fast enough. Brookings Institution. Available https://www.brookings.edu/blog/future-development/2019/03/28/pov erty-in-africa-is-now-falling-but-not-fast-enough/ International Finance Corporation (IFC) (2019) Executive summary, digital skills in Sub-Saharan Africa. Available https://www.ifc.org/wps/wcm/connect/58f4396b-fcee-49c8-82a4614fd3d53ea3/Digital+Skills+Report_WEB_ES.pdf?MOD¼AJPERES&CVID¼mGkda3h International Telecommunication Union (ITU) (2018) Measuring the information society report, Vol 2. Available https://www.itu.int/en/ITU-D/Statistics/Pages/publications/misr2018.aspx James J (2020) The smart feature phone revolution in developing countries: Bringing the internet to the bottom of the pyramid. Inf Soc 36(4):226–235 Kandri S (2019) Africa’s future is bright–and digital, World Bank blogs. Available https://blogs. worldbank.org/digital-development/africas-future-bright-and-digital

62

5 Interregional and Intercountry Analysis of Mobile Internet Connectivity in. . .

Madden P, Kanos D (2020) Figures of the week: digital skills and the future of work in Africa. Brookings Institution. Available https://www.brookings.edu/blog/africa-in-focus/2020/07/22/ figures-of-the-week-digital-skills-and-the-future-of-work-in-africa/ Michaelowa K (2001) Primary education quality in francophone Sub-Saharan Africa: determinants of learning achievement and efficiency considerations. World Dev 29(10):1699–1716 Ovide S (2019) The next big phones could bring a billion people online. Bloomsburg Businessweek, June 7. Available https://www.bloomberg.com/news/features/2019-06-07/thenext-big-phones-could-bring-a-billion-people-online Silver L, Johnson C (2018) Majorities in Sub-Saharan Africa own mobile phone, but smartphone adoption is modest. PEW Research Center. Available https://www.pewresearch.org/global/ 2018/10/09/majorities-in-sub-saharan-africa-own-mobile-phones-but-smartphone-adoption-ismodest/ Statista (2020) Gross domestic product per capita in selected global prices. Available https://www. statista.com/statistics/256413/gross-domestic-product-per-capita-in-selected-global-regions/ Statista (2021) Average price for mobile data in select African countries. Available https://www. statista.com/statistics/1180939/average-price-for-mobile-data-in-africa/ The Borgen Project (2020) 7 Facts about technology in Kenya. Available https://borgenproject.org/ technology-in-kenya/ UNESCO (2017) More than one-half of children and adolescents are not learning worldwide, Fact sheet No. 46. Available https://bangkok.unesco.org/content/more-one-half-children-and-adoles cents-are-not-learning-worldwide World Bank (2018) PovcalNet, Regional aggregation using 2011 PPP and $1.9/day poverty line, License CC BY-4.0 World Bank (2019) Future of food: harnessing digital technologies to improve food system outcomes. Available https://openknowledge.worldbank.org/handle/10986/31565 License: CC BY 3.0 IGO World Bank (2020) Data on youth literacy rate. Available https://data.worldbank.org/indicator/SE. ADT.1524.LT.ZS World Economic Forum (2019) Global competitiveness report. Available http://www3.weforum. org/docs/WEF_TheGlobalCompetitivenessReport2019.pdf World Economic Forum (2020a) Global competitiveness report. Available https://www.weforum. org/reports/the-global-competitiveness-report-2020 World Economic Forum (2020b) The future of the African economy. Available https://www. weforum.org/agenda/2020/02/africa-global-growth-economics-worldwide-gdp/

Chapter 6

Mobile Use of the Internet Among the Poor in the Global South: Preferences, Theories, and Policies

6.1

Introduction

Interest in the impact of the mobile Internet on the poor has grown rapidly in recent years, partly because devices such as smart feature phones and smartphones have spread widely in low- and low-middle-income countries over the same period (IDC 2021). In India, for example, there were 351 million users of mobile Internet in 2017, a number which is expected to grow to 420 million by 2023 (Statista 2019a). Interest has been heightened too, by the role played by the mobile Internet during the ongoing COVID-19 pandemic (Grantz et al. 2020). And finally, much interest has been sparked by the finding in at least two cases (see below) that the mobile device is used by poor individuals, mainly for the purpose of entertainment, rather than for traditional development goals, which would, hopefully, improve the health, productivity, and financial well-being of those involved. It is to this third issue that the chapter is mainly devoted and more specifically to the well-known anthropological interpretation of it that has been provided by Payal Arora (2016). I begin with a summary of her argument (Sect. 6.2), followed by a discussion of the evidence in support of it (Sect. 6.3). Then, I present a discussion of the alternative explanations (in Sect. 6.4), assuming, in spite of the somewhat limited evidence, that the poor do in fact spend more time on leisure and entertainment than those with higher incomes. Sections 6.5 and 6.6 deal, respectively, with the way skills are acquired in Africa and welfare implications more generally. Section 6.7 deals with the policy implications of my analysis, and these are followed in Sect. 6.8 by my conclusions.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. James, New Perspectives on Current Development Policy, SpringerBriefs in Economics, https://doi.org/10.1007/978-3-030-88497-0_6

63

6 Mobile Use of the Internet Among the Poor in the Global South:. . .

64

6.2

The Argument

Very near the beginning of her book Arora (2016) states her view of the reaction by development agencies to the growth of mobile Internet noted above. As she sees it, this [growth] has excited development agencies, which see in this digital network new opportunities to tackle poverty . . . Agencies have called upon Silicon Valley to produce applications that will offer, or improve access to jobs, health care, and education, and other public services [for the poor]. Their work is driven by the assumption that the poor will budget scarce digital resources and limited time online for seeking this information rather than for entertainment. Their attitude is fueled by a deep-seated worldview of the poor as utility-driven beings (Arora 2016, p. 1)

This view of the poor, she feels, runs against the vast evidence on internet users in general. Statistics on browsing patterns confirm that the sites most frequented online, whether in a suburb in Ohio, or a favela in Brazil, are social networking sites, pornography sites, romance sites, and gaming sites. People enjoy entertainment, romance, gaming and sex, regardless of their economic status. Although this is a readily accepted fact of contemporary digital life in the West, many still cling to the belief that the global poor are inherently different from typical users. Poverty, many assume, is a compelling enough reason for the poor to choose work over play when they go online (Arora 2016, p. 2).

Before going further into Arora’s views, it is worth noting here that there are a number of misstatements and inaccurate characterizations in the last quotation. For one thing, there is certainly no ‘vast evidence’ in the way poor individuals in the Global South behave on the Internet. On the contrary, as shown below, there are only a handful of empirical studies that directly address the issue.1 Then, too, it is doubtful that many people view the poor as being ‘inherently’ different from more affluent groups. The point is rather that people believe that the poor live under vastly different conditions than the typical Westerner. It is this, surely, that underlies the view that the poorest groups will pay more attention to Internet sites that deal with hunger, health, work, and so on. Then, too, the choice between work and leisure is presented as a binary one: the poor are said to choose between work and play. In reality, of course, not only are the lines blurred between these categories, but the choice is between how much of each will be chosen. Finally, and relatedly, the surprise that is said to be exhibited by development agencies and others is not so much occasioned by the choice of leisure on the part of the poor, as it is by how much time is spent on the activities that fall under this heading. Arora, unfortunately, does not deal with this issue, though it is, to my mind, a central part of a full analysis and one to which I will return. Further on in the prologue, Arora asks ‘Why should it even be a question whether people who are poor should enjoy themselves? Why do some people begrudge others who are struggling when they seek an occasional indulgence? Aren’t we all entitled to moments of pleasure and joy? Does poverty have to be miserable?’ 1

In her 2016 book, Arora cites only one article in her discussion of this topic.

6.2 The Argument

65

(emphasis added). The problem with these questions, however, is not that they are unworthy of answers, but rather that they tend to conflict with other claims made by Arora regarding Internet use by the poor. Consider first the reference to ‘moments of pleasure’ and ‘occasional indulgence’. These suggest that such experiences are fleeting rather than commonplace, an exception rather than a rule. Yet, at a later part of the book, one finds the claim that ‘leisure continues to be a central motivating force behind low-income communities’ adoption and use of digital media’ (Arora 2016, p. 14). Or again, she regards ‘differences in the access, intent, and use of digital leisure time as the dominant paradigm shaping global internet usage. The leisure economy needs a new description in terms of digital life today, especially among those, at the margins in the Global South’ (Arora 2016, p. 8). Although some readers may not make too much of these conflicting views of the role played by leisure in the Internet behaviour of the poor, they are in fact crucial to judging the welfare impact thus occasioned. For, while it is one thing to experience the fleeting bouts of happiness described by Arora, it is an altogether different story if leisure is pursued for long periods every day at the expense, say, of homework, or work on the family farm, not to speak of the deleterious psychological psycho-social effects that may then arise. In the Philippines, for example, where gaming is especially pronounced (Statista 2021), problems of addiction have been related to mental health disorders. One study of the issue, for instance, concludes that ‘Depression, as associated with online game addiction, is a serious threat that needs to be addressed. High level of online game addiction, as positively correlated to the rate of depression among adolescents in Manila, could potentially be attributed to the booming internet industry and lack of sufficient mental health interventions in the country’ (Labana et al. 2020, n.p.). In fact, gaming addiction has now been recognized by the American Psychiatric Association in its manual, DSM-5. And the World Health Organization has added ‘gaming disorder’ as an official mental health disorder, and it is included in the International Statistical Classification of Diseases and Related Health Problems (WHO 2018). More generally, what are needed to assess the welfare impact of leisure spending are data at the level of the poor individual, or group of individuals, which provide evidence on how much time is actually spent on social media, gaming, and other entertainment activities on the Internet. But as shown in Sect. 6.3, much of it is conducted at the level of countries and country comparisons, rather than individuals, and as such provides only suggestive evidence on the Arora thesis.2 My other main problem with her argument about the universality of preferences and the human need for leisure is that it largely ignores the possibility of other reasons for what she observes, especially those that lie outside her own discipline, anthropology. Such reasons, I suggest, lead to different policy conclusions than those that derive (implicitly) from Arora’s.

2

The point is that countries contain not only poor citizens but also some that are relatively wealthy.

66

6.3

6 Mobile Use of the Internet Among the Poor in the Global South:. . .

The Evidence

As just noted, much of the research on Internet use patterns in the Global South has taken place at the level of countries and comparisons between them. A study of poor countries by the GSMA (2019), for example, showed that among the least popular uses of the mobile Internet were those that could be described as developmental, such as gathering information for education, improving health, and government services. Social networking, on the other hand, was the second most popular activity. Similarly, research undertaken on African countries by the Research Institute Africa (RIA) showed that ‘More than half (55%) of people using the Internet in the surveyed countries, excluding Rwanda, spend most of their time on social media. Only 21% of users access educational content on the Internet, while 15% use the Internet for work’ (RIA 2019, p. 22). Most recently, a detailed study of Nigeria was undertaken by Yang Wang (2020). Although he divides his sample into various categories, he does not, unfortunately, classify them by income. Nor does he provide data on the amount of time and money spent on different Internet uses. Wang does, though, record a finding that will prove to be highly relevant to our discussion of alternatives to Arora’s view, namely that ‘Limited experience with the internet often results in a narrow view of all it has to offer. We found that both urban and rural respondents who claimed to use the internet regularly still stuck to very basic online activities’ (Wang 2020, n.p.). Data on Internet use choices by income group, which are needed to examine Arora’s thesis, in the context of the Global South, however, are regrettably scarce. One of the few studies that I could find is cited, but not discussed, by her. It comprises an examination of Internet use patterns in Nanjing, China, according to various socio-economic variables including income (in a regression analysis). The authors find among other things that ‘Low-income residents spent more time in the online entertainment activities’ (Chang et al. 2016, p. 57). On the basis of just this and a similar result for a developed country, however, Arora draws some sweeping conclusions. Thus, she finds that such evidence ‘is remarkable, as it turns the leisure economy on its head. The idea that the global poor would use their limited resources for entertainment instead of productive tasks is not good news for development agencies and governments’ (Arora 2016, p. 15). However, as argued below the welfare and policy implications of the preference of entertainment over work depend very much on why the phenomenon occurs and the time spent on it. As regards the former question (why it occurs), Arora has little to offer beyond the contention that ‘fulfillment is not necessarily a matter of efficiency or economic benefit but can involve a more elusive, personal and emotive drive’ (Arora 2016, p. 15). What she does not take into account, moreover, are the implications of the fact that in most poor countries, especially in sub-Saharan Africa, there is typically a relatively large family size (Statista 2019b) with a high percentage of young children. And while it is one thing for adults and older children to favour leisure over work activities on the Internet, this may often have implications for those too

6.3 The Evidence

67

young to register a preference, and who may, consequently, continue to suffer from a shortage of basic physical necessities. That is, that the opportunity cost of the pursuit of leisure does not fall only on those who make the choice. In cases where such costs are high, the role of development agencies is partly to protect those who cannot fend for themselves when physical needs are squeezed in the arguably excessive pursuit of leisure by other family members. Finally, the most recent study of Internet use by the poor is of Tanzania by Malm and Toyama (2021), which reports on mobile phone impact based on ‘133 semistructural interviews with low to middle-income individuals in Dar es Salaam and Arusha Region’ (Malm and Toyama 2021). But though it seems able to throw direct light on the question I am seeking to answer, it does not, in fact, add much to the evidentiary base. A major problem is that while some of the participants use non-connected feature phones, others rely on smartphones. One is thus comparing two very different groups from the point of view of Internet access. A related difficulty is that observed use patterns are not directly tied to income levels of the participants. What the authors do instead is to classify the sample by the type of mobile phone that they use. The group labelled ‘advanced’ is presumably the one that uses smart as opposed to feature phones and uses social media applications for marketing and e-commerce. What is not clear, however, is the income level of those who belong to this group or the share of time that they devote to particular Internet activities. The latter is important because the longer is the time spent on non-income generating activities, such as entertainment, the greater is the possibility that the basic physical needs of the poor family will be sacrificed (a phenomenon that was observed in the study of Tanzania referred to above by Malm and Toyama (2021).3 Moreover, beyond a certain point, time spent on entertainment may give rise to a variety of socio-psychological problems, such as isolation, depression, and addiction (referred to above). One needs to stress, in this regard, that as shown in Table 6.1, the top ten countries with the most time spent on social media per day are not the richest ones. Rather, they are drawn from what the World Bank refers to as lower- and upper-middleincome countries. For example, the average daily time spent by urban users on social media networks in 2020 was just over 4 h in the Philippines and 3 h 46 min in Nigeria (Global Web Index 2021). In fact, Table 6.1 shows that there are three other countries from sub-Saharan Africa in the top ten, namely, Kenya, South Africa, and Ghana. Interestingly, though, PEW shows that some of these countries also stand out in terms of information gathering uses on the Internet (PEW 2018). In particular, ‘Kenya and Nigeria are notable as two countries in which more internet users are going online to get information: In both countries, around six-in-ten internet users

This is the first intimation of the argument that a lack of digital skills is what drives poor users to choose relatively undemanding Internet uses, which, more often than not, are from entertainment rather than work. 3

68

6 Mobile Use of the Internet Among the Poor in the Global South:. . .

Table 6.1 Top ten countries, ranked by average daily time spent using social media according to urban Internet users, 2020 (hrs:mins)

Rank/country 1. Philippines 2. Nigeria 3. Brazil 4. Colombia 5. Kenya 6. South Africa 7. Mexico 8. Argentina 9. Ghana 10. Indonesia Average

Time 4:05 3:46 3:41 3.40 3:33 3:32 3:25 3:23 3:21 3:19 3:40

Source Global Web Index (2021)

say they go online to get political or government services information’ (PEW 2018, n.p.). Note that by comparison the world average in 2019 was 2 h and 15 min and some rich countries were in the vicinity of just one hour (Global Web Index 2021). Note, too, that four countries from sub-Saharan Africa are included in the table. Given the relatively intense use of social media by low- and middle-income countries, it is important that the welfare impact of this pattern be assessed. For, although it is broadly consistent with Arora’s thesis, she makes no attempt to delineate what might be considered as a legitimate amount of time spent on social media in a developing country context, where there are many competing options for spending time on the Internet, as noted above. Indeed, Arora does not pay more than passing attention to the whole issue of time spent on the media by the poor, or, for that matter, other groups as well. Before delving further into that crucial issue, however, it is well to first sum up this section on evidence in support of Arora’s thesis. The main finding is that while there are a few micro studies in support of her argument, there are others which seem on the face of them to be relevant, but which suffer from various methodological problems and ambiguities. At the country level, there is evidence that those in the Global South spend much more time on social media than those in the North.4 So much so, in fact, that questions surely arise about whether such time is excessive from the point of view of individual and family welfare in a poor country setting (after all, some such countries spend an average of around four hours per day on this type of Internet use). Indeed, there is already some evidence that the pursuit of leisure on the Internet leads to a reduction in the satisfaction of basic physical needs (Malm and Toyama 2021).5 What I have discussed above, however, is not the only serious criticism that can be levelled against the Arora formulation. For, even if further evidence confirms the 4 5

See Gillwald and Mothobi (2019). It was also observed in a study of durables in Brazil by Wells (1977).

6.4 Alternative Explanations

69

fact that the poor do spend more time on entertainment than higher income groups, there is still the question of whether alternative explanations exist for this finding. In the section that follows I suggest that this is indeed the case and that the welfare implications are very different from the inactivity that accompanies the assumption of consumer sovereignty that is made implicitly in the Arora scheme of things.

6.4

Alternative Explanations

It is worthwhile to emphasize at the outset of this section that the goal is not to explain what Arora refers to as ‘occasional indulgence’ or ‘moments of pleasure and joy’. Rather, as shown above, the use of social media in some countries of the Global South can run to up to four hours a day, a far cry from the fleeting moments of happiness that were just described. What is required instead is thus a more structural explanation of what occurs in reality. My proposal is based on the findings in a number of countries in the Global South that the most important barrier to effective Internet use is a lack of digital skills (GSMA 2021), which depends, among other things, on the extent and type of education the users receive. Stated briefly, my contention is that the pronounced lack of such skills in sub-Saharan Africa and elsewhere in the Global South largely precludes the use of developmental activities by the poor, which tend to be relatively demanding in comparison with leisure and entertainment.6 A shortage of digital skills is not an unknown problem in sub-Saharan Africa, but it is currently assuming critical proportion as the jobs of the future increasingly require a digital skills component. Thus, the IFC (2019) predicts that for this region, in excess of 230 million jobs will demand these skills by 2030. In fact, there are already reports that African firms are importing skilled labour from abroad in order to surmount the lack of local alternatives. Another way to view the shortage of digital skills in Africa is by comparing scores calculated for this variable by the World Economic Forum, for a selected sample of countries in the region, with a group drawn from the Global North. The results of this comparison are shown in Table 6.2. Although there are a few African countries that stand out from the rest, such as Kenya and Ghana with scores above 4 (out of 7), on the whole the divide in digital skills between the two groups is very stark in terms of absolute scores. In particular, compared to the African average of 3.5, the countries from the Global North achieve an average of 5.2. To this point, finally, I have implicitly assumed that developmental uses are available everywhere. But this may not be always the case and provides, to this extent, another reason why such choices are rarely made in the Global South.

Recall the finding reported in the text that first-time users of the Internet tend to turn to ‘shallow’ uses of it.

6

6 Mobile Use of the Internet Among the Poor in the Global South:. . .

70

Table 6.2 Digital skills scores in sub-Saharan Africa and developed countries, selected samples, 2019 (7 is highest) Sub-Saharan Africa Angola Burkina Faso Cameroon DRC Côte d’Ivoire Ghana Kenya Lesotho Malawi Mali Nigeria Senegal South Africa Average

Scores 2.45 2.89 3.9 2.8 3.8 4.21 4.55 3.49 3.6 3.6 3.42 4.21 3.27 3.5

Developed countries Denmark Finland France Germany Austria Hong Kong Japan The Netherlands Norway Singapore Switzerland UK USA Average

Scores 5.4 5.8 4.5 5.1 4.8 5.7 4.4 5.6 5.3 5.6 5.5 4.9 5.3 5.2

Source World Economic Forum (2019)

Unfortunately, however, this possibility has not, so far as I am aware, been examined, though it might turn out to be important in some circumstances.

6.5

Digital Skills, Information, and the Choice of Leisure

I turn now to examine my claim that the way in which digital skills are acquired in sub-Saharan Africa is a major reason why leisure rather than development-oriented uses are selected.7 Consider, in this regard, a study by Wang (2020) of first-time Internet users in Nigeria. He found overall that ‘respondents were interested in owning a phone and having mobile internet, but were not fully aware of all the benefits and ways to use the internet. . . . New users tend to learn about the internet from members of their local communities in an offline setting, which often leads to a narrow view of what the internet has to offer. For example, they might only hear about a few specific apps and have misconceptions about both the internet and how to access it’ (Wang 2020, n.p.). In effect, what Wang is arguing is that first-time users are lacking the digital skills needed to make more demanding uses of the Internet and, in consequence, turn to less demanding alternatives such as entertainment. Another study, by the GSMA, comes to much the same conclusion, arguing that ‘New users have made the leap into mobile Internet use, but usage is often shallow

7

Sub-Saharan Africa is selected because it has the lowest level of digital skills on average of all regions and may thus have the greatest predilection for leisure rather than development uses.

6.6 Welfare Implications

71

and restricted to a few key applications’ (GSMA 2018, p. 36). Shallow in this sense refers to activities that do not require complex digital skills and instead are rather superficial and basic, as is much of the Internet use devoted to entertainment in the Global South.

6.6

Welfare Implications

It is apparent from the aforementioned that first-time Internet users in poor countries tend to suffer from an acute lack of useful information of all kinds, having to rely instead on family and friends who may not be much better informed of the available possibilities. The issue is of considerable moment in a welfare context, because there are well-known economists such as Harsanyi (1992) who reject ill-informed preferences as a basis for normative economics. As he puts it, ‘for normative issues informed preferences should be used, instead of actual preferences or happiness (or welfare)’ (Ng 1999). Examples of where uninformed preferences are rejected are easy to find in the Global North and include water fluoridation, banning of dangerous drugs, enforcement of mask-wearing, compulsory pensions, and certain insurance policies. In the South, by contrast, and especially in the least developed countries, misinformation and disinformation are rife, partly because of a lack of countervailing policies and partly because of an almost unfettered influence over information exerted by multinational corporations (James 1983). In the case of medicinal drugs, for example, numerous books attest to the way in which patient information is systematically less accurate in the Global South compared to the North.8 How then should one treat the apparently extensive reliance on entertainment and social media by at least some of those living in poverty in the Global South? One important question, as noted above, is the knowledge about the various use options that are potentially available. The material I cited above suggests that such knowledge tends on the whole to be severely lacking in sub-Saharan Africa. The preferences exhibited there, therefore, do not nearly pass the test of being informed and as such cannot be defended by the notion of ‘consumer sovereignty’ that plays a central role in traditional micro-economics (Hyman 1989). This is because consumers are then regarded as sacrosanct and their preferences are thus not to be challenged, except, perhaps, under extreme circumstances.9 Yet, economists and psychologists who have studied human decision-making repeatedly find that consumers are not always the best judges of their own welfare:

8

One of them is by Silverman et al. (1992), in the context of medicinal drugs supplied by multinationals to poor countries. The evidence is very compelling. 9 Consumer sovereignty lies at the heart of non-interventionist policy prescriptions in general, including policy towards consumption in the Global South.

72

6 Mobile Use of the Internet Among the Poor in the Global South:. . .

that they are often irrational and subject to certain biases in their decision-making (Thaler 1992). That is part of why there are so many interventions in their behaviour in the Global North, covering insurance, health, traffic, education, and so on. In the South, however, the regulatory machinery is usually grossly underdeveloped and policies that override consumer sovereignty are typically much less in evidence (with respect, for example, to health, medicine, automobiles, seatbelts, and counterfeit products). For what Arora at times describes as the fleeting pleasure derived by the use of the Internet for entertainment purposes, there is probably not much to be concerned about from a welfare point of view. Much larger issues, however, emerge when the daily use of social media runs to three or four hours (see Table 6.1). These cases are unfortunately not raised by Arora, though at that level of use, serious questions arise as to what constitutes excessive use, as when, for example, social isolation and a deterioration of school performance begin to set in, not to speak of addictions to gaming and pornography, The occurrence of such negative effects is hardly surprising in the light of the opinion by a group of psychologists that ‘our findings strongly suggest that limiting social media use to approximately 30 min per day may lead to significant improvement in well-being (Hunt et al. 2018, emphasis added). Although one can quibble with this exact amount of time, it nevertheless provides us with a rough idea of the excessive degree of use in many countries in the Global South.

6.7

Policy Implications

In an Arora-like world, user preferences are unquestionably accepted on the basis of consumer sovereignty, which assumes, among other things, that there is perfect information and that each Internet user is the best judge of her own well-being. In such a world, economic policy tends towards non-intervention or laissez-faire, such as in the case of free trade. In reality, however, and especially in the Global South, the notion of consumer sovereignty is open to wide criticism and so too therefore is the policy with which it is associated. In many countries in that region, for example, first-time Internet users appear to be very poorly informed about what the Internet offers and how it works, because they do not have access to sources of reliable information. Even more telling is the discrepancy between the average daily time spent on social media by some countries in the South, of three or four hours per day, and the recommendation by psychologists that 30 minutes is the maximum time that should be spent (Hunt et al. 2018). Excessive use of social media on the mobile Internet is known to cause a host of psycho-social problems, ranging from social isolation to depression and addiction to gaming and pornography. In such cases, intervention rather than laissez-faire seems the appropriate policy response. Then, too, I argued that Arora pays no attention to alternative explanations of the seemingly anomalous behaviour of first-time Internet users in the South. Several studies cited above, for example, attributed the emphasis of first-time users on leisure

6.7 Policy Implications

73

rather than work to a lack of digital knowledge, which resulted in use choices that were undemanding in terms of that knowledge, which meant for the most part a preference for entertainment. According to this view, the said preference is a derived one, rather than being some innate preference for very large amounts of entertainment. If this is indeed the prevailing pattern in the Global South, much more needs to be done to promote levels of digital skills, especially among first-time users of the Internet. Policy of this kind is also sorely needed at the macro level, because of the widening gap referred to above, between the demand for these skills and the supply of them in the poor country as a whole. It is not my intention here, though, to review all the efforts that are currently being made to redress the shortage of digital skills in the Global South. Rather, I focus on first-time users of the Internet, who have not been able to acquire these skills through learning by doing. As the Nigeria case showed, such persons are especially prone to rely on those close to them, who, themselves, hold only a very limited view of what the Internet has to offer, and how to operate it. One prominent means of overcoming or ameliorating this type of problem is an application called ‘Life’, designed by KaiOS Technologies, which provides reliable information on a variety of development issues. And there is some evidence that it has become popular in sub-Saharan Africa. In particular, ‘Life is now among the top five most downloaded apps in Nigeria and Uganda, and ranks 14th in Rwanda. Users in these countries can access a directory of free resources curated for first-time internet users to improve everything from their digital literacy to their health’ (Mak 2020, n.p.). With regard to the training in digital skills, Life offers tutorials in which users lacking most such skills are taught to use tools such as Google Maps, WhatsApp, and YouTube. ‘Modules are designed not only to educate consumers about how to use certain apps but also why. For example, consumers will learn how to save on SMS and voice call fees by using WhatsApp for communications. Other digital skills modules connect users with trustworthy news sources, and educate consumers about internet scams’ (Mak 2020, n.p.). ‘For the newly connected’, writes Meta (2019), ‘these materials are invaluable’. What is not yet known about this endeavour, however, is whether and to what extent it causes a shift away from entertainment uses to those that are more developmental. The same question applies, moreover, to other attempts to bring more accurate information about the Internet to first-time users. Unfortunately, however, not only is it unclear whether digital skills are actually fostered, but also, if so, whether they bring about a more developmental pattern of Internet use. It is unfortunate because the answers to these questions are highly important to the formulation of policies to raise the level of digital skills among first-time users of the Internet. Research on these issues is thus clearly indicated.10

10

One issue then is whether and under what circumstances digital skills are increased and another is whether, if so, a more developmental pattern of Internet use is the result.

6 Mobile Use of the Internet Among the Poor in the Global South:. . .

74

It is also needed in the case of large-scale national programmes such as that initiated in 2017 by the Indian Prime Minister. The scheme is designed to bring digital skills to 60 million people in rural areas, which would involve 40% of rural households, by including one family member from every eligible household. The Scheme would empower the citizens in rural areas by training them to operate computer or digital access devices . . . send and receive e-mails, browse Internet, access Government services, search for information, undertake digital payment etc. and hence enable them to use the Information Technology and related applications especially Digital Payments . . . The Scheme aims to bridge the digital divide, especially targeting the rural population including the marginalized sections of society like Scheduled Castes . . . Minorities, Below Poverty Line (BPL), women and differently-abled persons’ (PMGDISHA 2021.)

Apparently because of a shortage of funds, however, after three years the scheme had managed to train only a portion of the set target. This still represents a large absolute number of beneficiaries, and it would be valuable to study the extent to which digital skills have increased as a result of the programme.

6.8

Conclusions

There is some evidence in favour of Payal Arora’s thesis that Internet use among the poor in the Global South is dominated by leisure, in contrast to the North where more development-oriented uses hold sway. What she does not take into account, however, is the amount of time spent by the poor in the South, which, in my view, carries certain major welfare implications. One of them is that the greater is this amount, the less it can be defended on the grounds that it serves to meet the ‘moments of pleasure’ and ‘occasional indulgence’ mentioned by Arora. Consider in this regard that in a number of countries in the South, the average daily amount of time spent on social media is around three to four hours. Another implication is that at these amounts of time spent, questions surely need to be asked about excessive use, as when users begin to suffer from a variety of negative socio-psychological effects, such as social isolation, depression, and addiction to gaming and pornography. The prevalence of such effects (in certain countries) is hardly surprising when one compares the three or four hours mentioned above with the maximum of thirty minutes per day recommended by experts, as noted in the foregoing text. My other main problem is also methodological and has to do with her failure to consider alternative explanations of her findings. That is, that she does not venture out of her anthropological perspective, although the issue at hand seems to require it. My argument, for example, turns on the notion of digital skills, which are rarely linked to Arora’s discipline. What I propose begins with the acute scarcity of such skills in a very poor region such as sub-Saharan Africa, as described above. Faced with such a severe disadvantage, my suggestion is that such consumers are forced to rely on Internet uses that are undemanding in terms of these skills. For the most part, this means entertainment

References

75

rather than developmental alternatives. In this argument, therefore, the demand for leisure is a derived one, based on severe limitations of digital skills. If this line of argument has merit, then the appropriate policy response is not the inactivity and reliance on the status quo that accompany the notions of consumer sovereignty and laissez-faire, but rather a host of attempts to raise digital skills, several of which were provided in the text. And it is of the utmost importance that these and other efforts enquire whether any increases in digital skills are accompanied by a move away from entertainment to more developmental activities.11 If not, at least the possible squeeze on basic nutritional needs among the poor referred to above can be ameliorated by food fortification. Banerjee et al. (2018), for example, have investigated the possibility of reducing iron deficiency anaemia by using salt fortified with iron and iodine.

References Arora PA (2016) The next billion users: digital life beyond the West. Harvard University Press Banerjee A, Barnhardt S, Duflo E (2018) Can iron-fortified salt control anemia? Evidence from two experiments in rural Bihar, J Dev Econ 133:127–146 Chang E, Zhen F, Cao Y (2016) Empirical analysis of the digital divide from the perspective of internet usage patterns: a case study of Nanjing. Int Rev Spat Plan Sust Dev 4(1):49–63 Gillwald A, Mothobi O (2019) After access 2018: a demand-side view of mobile internet from 10 African countries. Research ICT Africa, Policy Paper No. 7. Available https://www. africaportal.org/publications/after-access-2018-demand-side-view-mobile-internet-10-africancountries/ Global Web Index (GWI) (2021) Social flagship report. Available https://www.emarketer.com/ content/top-10-countries-where-people-spend-most-time-on-social-media Grantz KH, Meredith HR, Cummings DAT et al (2020) The use of mobile phone data to inform analysis of COVID-19 pandemic epidemiology. Nat Commun 11:4961 GSMA (2018) Triggering mobile internet use in Côte d’Ivoire and Tanzania. Available https:// www.gsma.com/mobilefordevelopment/resources/triggering-mobile-internet-use-in-cotedivoire-and-tanzania/ GSMA (2019) The state of mobile internet connectivity report. Available https://www.gsma.com/ mobilefordevelopment/resources/the-state-of-mobile-internet-connectivity-report-2019/ GSMA (2021) Understanding people’s mobile digital skill needs. Available https://www.gsma. com/mobilefordevelopment/resources/understanding-peoples-mobile-digital-skills-needs/ Harsanyi J (1992) Utilities, preferences and substantive goods. UNU WIDER working paper 101. Available https://www.wider.unu.edu/sites/default/files/WP101.pdf Hunt M, Marx R, Lipson C, Young J (2018) No more FOMO: limiting social science media decreases loneliness and depression. J Soc Clin Psychol 37(10):751–768 Hyman DN (1989) Microeconomics. Irwin Professional Publishing International Data Corporation (IDC) (2021) Smartphone growth to reach its highest level since 2015. Available https://www.idc.com/getdoc.jsp?containerId¼prUS47770921

11

An obvious place to focus on digital skills training is schools, where currently the topic receives a disappointing amount of attention, perhaps especially in sub-Saharan Africa. According to a World Bank blog written in 2019, ‘only 50% of countries in Africa have computer skills as part of their school curriculum compared to 85% of countries globally’ (World Bank 2019).

76

6 Mobile Use of the Internet Among the Poor in the Global South:. . .

International Finance Corporation (IFC) (2019) Digital skills in Sub-Saharan Africa: spotlight on Ghana. Available https://www.ifc.org/wps/wcm/connect/ed6362b3-aa34-42ac-ae9fc739904951b1/Digital+Skills_Final_WEB_5-7-19.pdf?MOD¼AJPERES James J (1983) Consumer choice in the third world: a study of the welfare effects of advertising and new products in a developing country. Macmillan Labana RV, Hadjisaid JL, Imperial AR, Jumawid KE, Lupague MJM, Malicdem DC (2020) Online game addiction and the level of depression among adolescents in Manila, Philippines. Cent Asian J Glob Health 9(1) Mak I (2020) How the Life app is changing lives around the world. Available https://www. kaiostech.com/life-app-changing-lives-around-world/ Malm MK, Toyama K (2021) The burdens and the benefits: socio-economic impacts of mobile phone ownership in Tanzania. World Dev Perspect 21:100283 Meta A (2019) Life, the KaiOS in-house app that bridges the gap to educational resources, is now live in Africa. Available https://www.kaiostech.com/life-the-kaios-in-house-app-that-bridgesthe-gap-to-educational-resources-is-now-live-in-africa/ Ng Y-K (1999) Utility, informed preference, or happiness: following Harsanyi’s argument to its logical conclusion. Soc Choice Welfare 16:197–216 PMGDISHA (2021) 30 million registered beneficiaries under PMGDISHA. Available https:// twitter.com/PMGDISHA Research ICT Africa (RIA) (2019) After access: Africa comparative report. Available https:// researchictafrica.net/2019/05/16/after-access-africa-comparative-report/ Silverman M, Lydecker M, Lee PR (1992) Bad medicine: the prescription drug industry in the third world. Stanford University Press, Stanford CA Statista (2019a) Number of mobile phone Internet users in India from 2015 to 2018. Available https://www.statista.com/statistics/558610/number-of-mobile-internet-user-in-india/ Statista (2019b) Average household size worldwide in 2019, by region. Available https://www. statista.com/statistics/1090668/average-household-size-worldwide-by-region/ Statista (2021) Duration of one session of online gaming in the Philippines as of March 2020. Available https://www.statista.com/statistics/1117439/philippines-duration-of-one-gaming-ses sion-online-gamers/ Thaler R (1992) The winner’s curse: paradox and anomalies in economic life. Free Press Wang Y (2020) First-time internet users in Nigeria use the internet in a unique way. Here’s why that matters. Available https://www.kaiostech.com/first-time-internet-users-in-nigeria-use-the-inter net-in-a-unique-way-heres-why-that-matters/ Wells J (1977) The diffusion of durables in Brazil and its implications for recent controversies concerning Brazilian development. Camb J Econ 1(3):259–279 World Bank (2019) Africa’s future is bright and digital. World Bank blogs. Available https://blogs. worldbank.org/digital-development/africas-future-bright-and-digital World Economic Forum (2019) The global competitiveness report. Available http://www3. weforum.org/docs/WEF_TheGlobalCompetitivenessReport2019.pdf World Health Organization (WHO) (2018) Classification of diseases. Available https://www.who. int/standards/classifications/classification-of-diseases